CN101068323A

CN101068323A - Information processing device, information processing method, program and recording medium

Info

Publication number: CN101068323A
Application number: CNA2007101049744A
Authority: CN
Inventors: 山本则行; 宫崎充弘; 斋藤真里; 小池宏幸
Original assignee: Sony Corp
Current assignee: Sony Corp
Priority date: 2002-12-12
Filing date: 2003-12-12
Publication date: 2007-11-07
Anticipated expiration: 2023-12-12
Also published as: JP4003127B2; CN1736097A; CN101068325A; CN101068324B; CN100499761C; CN101068324A; CN100499756C; CN100486310C; JP2004194107A

Abstract

The present invention provides Information processing device, information processing method, program and recording medium wherein a program vector representing attributes of a program is generated as follows. At a step S 11 , EPG data is received. At a step S 12, meta data necessary for generation of a program vector PP is extracted from the EPG data. At a step S 13 , a morphological analysis is carried out on contents and title included in the meta data to disassemble the contents and the title into words. At a step S 14, items included in the meta data are subjected to a vector creation process to generate the program vector PP. At a step S 15, an effect vector is extracted on the basis of a genre of a program associated with the meta data. At a step S 16, the extracted effect vector is associated with the generated program vector PP and the processing is ended. The present invention can be applied to a distribution server for distributing contents.

Description

Messaging device, information processing method, program and recording medium

The application is that application number is 200380108320.9, the applying date is on December 12nd, 2003, denomination of invention is divided an application for the invention of " messaging device, information processing method, information processing system, recording medium and computer program ".

Technical field

The present invention relates to messaging device, information processing method, information processing system, recording medium and computer program.More particularly, the present invention relates to as the equipment that is suitable for the application such as the recommendation of automatic record and program, the messaging device that provides for the user of the stream data of watching program and television broadcasting and radio broadcasting usually, also relate to the information processing method that described equipment adopts, adopt the information processing system of this equipment, realize the computer program of described method, and the recording medium that is used for the logger computer program.

Background technology

By convention, in order to recommend the program in TV and the radio broadcasting to the user, select and the program of the information matches that the user likes by using programme information (perhaps program metadata) such as EPG (electronic program guides).Change along with the method that obtains the data that the user likes to the programs recommended method of user.An example to the programs recommended method of user is initial interest categorization, and in the method, the information relevant with user's interest is at first made a catalogue (catalog) in advance, and is used as to the programs recommended basis of user.Another example to the programs recommended method of user is to watch history to utilize method, and this method is by the history of the program that utilizes the user and watch in the past, and is programs recommended to the user.Another example to the programs recommended method of user is by using other user's viewing history, to the programs recommended collaborative filtering method of user.

In initial interest categorization, at first require the user in advance to kind, the information of name of type of liking and favorite actor (talent) and so on catalogue such as the program of liking.The example of the kind of the program of liking is drama and variety show.The example of the type of liking is medieval mystery and comedy.Subsequently, inventory information is used as the keyword of discerning the program metadata that conforms to user's interest.At last, acquisition will be to the title of user's recommend programs.

Utilize in the method watching history, when each user watches program, the metadata of the program of being watched is saved, when the metadata in past is saved to a certain degree, analyze the metadata of preserving, thereby obtain kind, the information of the type of liking and name of favorite actor and so on such as the program of liking.Subsequently, the information of acquisition is used as identification and user's the keyword of liking the program metadata that conforms to.At last, acquisition will be to the title of user's recommend programs.

In the equipment such as the recording equipment that adopts HDD (hard disk drive), not only according to watching history, and, obtain the information relevant with user preferences in some cases also according to the history of the historical of recording reservation or the user's operation such as recording operation.In this case, can be interested the accidental program of watching of user and user and the program area of having a mind to watch separately, thereby can obtain to reflect better the information of user's hobby.The accidental program of watching of user is not because the user is interested in especially and watch this program and the program that is provided with by the user program, but the program that the user who is presented by the television receiver that is in open mode (or broadcast receiver) watches.

In collaborative filtering method, first user at first searches for has watching of conforming to the watching of first user/operation history/second user of operation history, obtains watching/operation history of second user.Subsequently, the program of in the program that second user watched, selecting first user also not watch, and recommend first user.

In addition, the open No.2001-160955 of Japanese Patent Laid has proposed a kind of technology, thereby n dimension attribute vector is joined in the broadcast program in advance as the attribute of program.Relatively select the vector sum attribute vector subsequently, thereby select program that will write down or the program that will reproduce.Selecting vector is data according to the initial record of user, and the vector of the mean value generation of the attribute represented of the attribute vector of the program of program that is reproduced by the user or reservation recording.

But, if adopt initial interest categorization to select program, when the program of a Xuan Zeing representative of consumer is made a catalogue to information at first so, the special interests that this user has.In addition, for the details of recorded information, the user need carry out complicated information catalogue operation.Thereby,,, need to reduce the number of packages of the information that will write down to the operation that the information that will write down is made a catalogue in order to simplify in the initial set-up procedure.Thereby, be merely able to recommend the program of selecting according to the coarse information of user record.Thereby the accuracy of the program that selection conforms to user's hobby is lower.

On the other hand,, only utilize the summation or the mean value of many metadata of collecting, select recommend programs according to user's viewing history such as watching history to utilize in other method the method.Thereby in this case,, can not correctly recommend the program that conforms to user's hobby so if history is not saved to certain degree.In addition, utilize method with regard to watching history, the correlation between many metadata is also insensitive, causes realizing individualized fully.In addition, if historical the accumulation, so in some cases, because that history entries is easy to is overlapped, and history entries all has and expands to historical element easily, therefore may form deviation on weighted value.Overlapped history entries is the project that is detected as user's hobby easily easily.An example of overlapped history entries is a type easily.On the other hand, having the history entries that expands to historical element easily is the project that is difficult to be detected as user's hobby.An example with the history entries that expands to historical element easily is the performing artist.

More particularly, suppose that the user is the admirer of commentator A.Thereby this user likes watching the commentary of the baseball game of the B of team that commentator A explains orally.In this case, information " commentary of baseball game " (it is a type) is overlapped easily as history.That is, information " commentary of baseball game " is detected as user's hobby easily.But information " commentator A " (it is the performing artist) is difficult to overlapped.That is, information " commentator A " is difficult to be detected as user's hobby.Thereby existence is recommended by the commentary of the baseball game of the B of team of another commentator's explanation, and the not recommended situation of variety show of commentator A performance.

In addition, as the open No.2001-160955 of Japanese Patent Laid was disclosed, attribute vector was joined in the broadcast program in advance.Relatively select the vector sum attribute vector subsequently, thereby select program that will write down or the program that will reproduce.Selecting vector is data according to the initial record of user, and the vector of the mean value generation of the attribute represented of the attribute vector of the program of program that is reproduced by the user or reservation recording.In addition in this case, owing to use user's operation history, therefore in some cases, because that history entries is easy to is overlapped, and history entries all has and expands to historical element easily, such as the performing artist, therefore may form deviation on weighted value.

For example suppose that the user likes drama, and only like not performing the variety show of the comedian A of drama.In addition, suppose that it is 2: 8 that the user watches the ratio of such variety show and drama.In the selection vector that produces about such user, the information of " performing artist B " of often performing drama is overlapped as history, rather than performs the comedian A of drama hardly, although performing artist B is not the star of the special hobby of user.Thereby it is recommended as performing artist's variety show that the documentary film program that " the performing artist B " that often performs drama participates in performance has precedence over comedian A.

In addition, important project is different because of the user concerning the selection of program.For example, concerning a certain user, performing artist importantly, and concerning another user, the importantly content of program.Yet,, therefore in some cases, in recommend programs, can not reflect the hobby of having only the user just to have owing to operate all items according to identical mode.

In addition, because what utilize is another user's hobby in the collaborative filtering method, therefore be difficult to extract the information of detailed each user's of expression hobby.

Summary of the invention

In order to address the above problem, the present inventor has considered the ability of the program that selection conforms to user's hobby.

In order to address the above problem, the invention provides a kind of messaging device of carrying out the processing of chosen content, comprising: the deriving means that obtains the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of described deriving means acquisition and from the calculation of similarity degree device between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

The present invention also provides a kind of information processing method of messaging device of the processing of carrying out chosen content, comprising: the obtaining step that obtains the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of the processing acquisition of described obtaining step and from the calculation of similarity degree step between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

The present invention also provides a kind of and is used to make computer to carry out the program of the processing of chosen content, comprising: the obtaining step that obtains the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of the processing acquisition of described obtaining step and from the calculation of similarity degree step between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

The present invention also provides a kind of recording medium, has write down said procedure.

The present invention also provides a kind of messaging device of carrying out the processing of chosen content, comprising: first storage device of preserving the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of described first storage device preservation and from the calculation of similarity degree device between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

The present invention also provides a kind of information processing method of carrying out the processing of chosen content, comprising: first storing step of preserving the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of the processing preservation of described first storing step and from the calculation of similarity degree step between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

The present invention also provides a kind of and is used to make computer to carry out the program of the processing of chosen content, comprising: first storing step of preserving the attribute information that comprises a plurality of projects of described content; And utilize predetermined weighted value information calculations by the described attribute information of the processing preservation of described first storing step and from the calculation of similarity degree step between the information of user's acquisition, wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

According to the present invention, the information of a kind of basis about content is provided, produce the first information treatment facility of the attribute information of content, comprising:

Acquisition is about the deriving means of the information of content;

According to the information that deriving means obtains, produce the attribute information generation device of the attribute information that comprises a plurality of projects about content; With

First storage device of the first weighted value information of each project of the attribute information that preservation attribute information generation device produces,

Each percentage contribution in the described a plurality of projects of the first weighted value information specifies wherein to the calculation of similarity degree between attribute information and user's the predetermined preference information.

Best, messaging device also comprises the transmitting device of the attribute information that transmits the generation of attribute information generation device.

Best, extract one the first weighted value information that conforms to the condition of content in many first weighted value information of transmitting device from be kept at first storage device, and be associated with the attribute information that the attribute information generation device produces by one the first weighted value information that makes extraction, transmit one the first weighted value information that extracts.

Best, the condition of content is the type of content.

Best, messaging device also comprises the draw-out device about extraction predetermined information the information of content that obtains from deriving means, wherein attribute information generation device predetermined information that draw-out device is extracted converts each vector of a plurality of projects to, so that produce attribute information.

Best, the attribute information generation device from about the information selecting the information of content to represent with word as analytic target, and produce attribute information according to the result who analyzes.

Best, messaging device also comprises: second storage device that the information that comprises a plurality of projects is saved as user's predetermined preference information; With by about each project, calculate the similarity between the attribute information that the preference information that is kept in second storage device and attribute information generation device produce, produce the recommendation information generation device of the recommendation information of the content that announcement conforms to user's hobby.

Best, the recommendation information generation device is kept at the first weighted value information in first storage device by use, via the comparison of attribute information and preference information, produces recommendation information.

Best, messaging device also comprises:

The operation history deriving means of the history of the operation that the acquisition user carries out;

The history of the operation that the user carried out that obtains according to the operation history deriving means produces the preference information generation device of user's preference information; With

According to the preference information that the preference information generation device produces, produce the weighted value information generating apparatus of the second weighted value information, wherein:

The second weighted value information is indicated each percentage contribution to the calculation of similarity degree between attribute information and user's the preference information in a plurality of projects;

The second weighted value information of recommendation information generation device by using the weighted value information generating apparatus to produce via the comparison of attribute information and preference information, produces recommendation information.

According to the present invention, the information of a kind of basis about content is provided, produce the messaging device first information processing method of the attribute information of content, described information processing method comprises:

Acquisition is about the obtaining step of the information of content;

According to the information about content that obtains in the processing of carrying out at obtaining step, the attribute information that produces the attribute information that comprises a plurality of projects produces step;

According to the information that obtains in the processing of carrying out at obtaining step about content, and, extract each the extraction step in a plurality of projects of regulation to the weighted value information of the percentage contribution of the calculation of similarity degree between attribute information and user's the predetermined preference information according to the condition that content has; With

Make the weighted value information that in the processing that extraction step is carried out, extracts and produce the associated steps that the attribute information that produces in the processing that step carries out is associated at attribute information.

According to the present invention, provide a kind of be used to store by computer carry out so that carry out producing according to information about content first recording medium of computer program of processing of the attribute information of content, described computer program comprises:

Acquisition is about the obtaining step of the information of content;

According to the present invention, provide a kind of and carry out by computer, so that carry out producing according to information about content first computer program of processing of the attribute information of content, described computer program comprises:

Acquisition is about the obtaining step of the information of content;

As mentioned above, obtain information, and, produce the attribute information that comprises a plurality of projects according to the information that obtains about content about content.Subsequently, according to information about content, and, attribute information and regulation each in a plurality of aforementioned projects are associated to the weighted value information of the percentage contribution of the calculation of similarity degree between attribute information and user's the predetermined preference information according to the condition that content has.

According to the present invention, second messaging device of the processing of selecting the content that conforms to user's hobby is provided, comprising:

Acquisition is as the deriving means of the information that comprises a plurality of projects of the attribute information of content;

Preserve the storage device of the information that comprises a plurality of projects of the preference information that is used as the user;

By using predetermined weighted value information, calculating is kept at similarity between the attribute information that information in the storage device and deriving means obtain as user's preference information, produce the recommendation information generation device of the recommendation information that discloses the content that conforms to user's hobby

Each percentage contribution in a plurality of projects of weighted value information specifies wherein to the calculation of similarity degree between attribute information and user's the preference information.

Best, except the attribute information of content, deriving means also obtains weighted value information, and the recommendation information generation device is by utilizing weighted value information, and relatively the attribute information and the described preference information of content produce recommendation information.

Best, messaging device also comprises:

According to the preference information that the preference information generation device produces, produce the weighted value information generating apparatus of weighted value information,

Wherein the weighted value information of recommendation information generation device by utilizing the weighted value information generating apparatus to produce compares preference information and attribute information, produces recommendation information.

Best, weighted value information is to disclose the peculiar hobby of user's first-selection, rather than the information of general hobby, and described peculiar hobby is used to the specific attribute information in many attribute informations of chosen content.

Best, weighted value information is the information that discloses in the project of attribute information of constitution content an important project of user.

Best, weighted value information is to disclose in the project of attribute information of constitution content the information of the project of the content that the indication user likes.

Best, weighted value information is to disclose in the project of attribute information of constitution content the information of the project of the content that the indication user dislikes.

Best, messaging device also comprises the input device of reception from user's operation input, wherein according to the operation input of user's input operation input unit, weighted value information is set.

According to the present invention, a kind of messaging device second information processing method of selecting the processing of the content that conforms to user's hobby is provided, described information processing method comprises:

Acquisition is provided with the obtaining step of the information of weighted value information, described weighted value information specifies is to being made of a plurality of projects, as the information of the attribute information relevant with constitute by a plurality of projects with content, as and the information of user-dependent predetermined preference information between the percentage contribution of calculation of similarity degree, contribution described here is the contribution that is caused by each project in a plurality of projects;

According to the information that weighted value information is set that obtains in the processing of carrying out at obtaining step, the calculation of similarity degree step between computation attribute information and the preference information; With

By the use of the result of calculation that obtains in the processing of carrying out in calculation procedure, produce the recommendation information generation step of the recommendation information that discloses the content that conforms to user's hobby.

According to the present invention, provide a kind of preservation to carry out by computer, select second recording medium of computer program of the processing of the content that conforms to user's hobby, described computer program comprises:

According to the present invention, providing a kind of will be carried out by computer, select the computer program of the processing of the content that conforms to user's hobby, and described computer program comprises:

As mentioned above, in order to produce the recommendation information that discloses the content that conforms to user's hobby, by use each the weighted value information in a plurality of following projects of regulation, calculate the percentage contribution that calculates comprise a plurality of projects as the similarity between the information of the information of the attribute information of content and the predetermined preference information that is used as the user that comprises a plurality of projects.

According to the present invention, a kind of information processing system is provided, described information processing system comprises according to the first information treatment facility that produces the attribute information of content about the information of content, with attribute information according to the content that receives from first information treatment facility, select second messaging device of the processing of the content that conforms to user's hobby

First information treatment facility comprises:

Acquisition is about first deriving means of the information of content;

According to the information that first deriving means obtains, produce the attribute information generation device of the attribute information that comprises a plurality of projects about content;

First storage device of the first weighted value information of each project of the attribute information that preservation attribute information generation device produces; With

Extract one the first weighted value information that conforms to condition that content has in many first weighted value information from be kept at first storage device, the first weighted value information of extraction is associated with the attribute information that the attribute information generation device produces, transmitting device with the first weighted value information of transmission and attribute information

Second messaging device comprises:

Acquisition comprises the attribute information of content of a plurality of projects and second deriving means of the first weighted value information;

Preserve second storage device of the information that comprises a plurality of projects of the preference information that is used as the user; With

By using the first weighted value information at least or being different from the second weighted value information of the first weighted value information, calculating is kept at similarity between the attribute information that information in second storage device and second deriving means obtain as user's preference information, produce the recommendation information generation device of the recommendation information that discloses the content that conforms to user's hobby

Wherein the first weighted value information and the second weighted value information are all stipulated each percentage contribution to the calculation of similarity degree between attribute information and user's the preference information in a plurality of projects.

As mentioned above, in first information treatment facility, obtain information, and, produce the attribute information that comprises a plurality of projects according to the information that obtains about content about content.Subsequently, preserve the first weighted value information of each project of attribute information.Afterwards, from many first weighted value information of preserving, extract one the first weighted value information that conforms to condition that content has, and make it to be associated with attribute information.At last, the be mutually related first weighted value information and attribute information is transmitted to second messaging device.On the other hand, in second messaging device, from first information treatment facility, and be saved as the information that comprises a plurality of above-mentioned projects of user's preference information as the information that comprises a plurality of projects of the attribute information of content and the first weighted value message pick-up.Subsequently,, calculate the user's who preserves preference information and the similarity between the attribute information, produce the recommendation information that discloses the content that conforms to user's hobby by using the first weighted value information at least or being different from the second weighted value information of the first weighted value information.In addition, the first weighted value information and the second weighted value information are all stipulated each percentage contribution to the calculation of similarity degree between attribute information and user's the preference information in a plurality of above-mentioned projects.

According to the present invention, provides and handle, thereby the 3rd messaging device of the content that selection conforms to user's hobby comprises:

Obtain the deriving means of the attribute information of content; With

According to the first information of the hobby of representing the user and second information of the general hobby of expression, generation discloses the deviation information generation device of user's hobby with respect to the 3rd information of the deviation of general hobby.

Best, attribute information, the first information and the 3rd information include a plurality of projects; Be arranged on choice device in the messaging device in addition by utilizing the 3rd information, about each project, the similarity between the computation attribute information and the first information is selected the content that conforms to user's hobby.

Best, the first information and second information include a plurality of projects; The generation of deviation information generation device discloses the information of the project with the low similarity between the first information and second information as the 3rd information.

Best, messaging device also comprises: the operation history deriving means of the history of the operation that the acquisition user carries out; The history of the operation of carrying out with the user who obtains according to the operation history deriving means produces the preference information generation device of the first information.

Best, the deviation information generation device produces the 3rd information by following operation:

About each predetermined project, the user is subordinated to the number of contents counting of selecting in the content of being scheduled to group and watching, thereby obtains first value, and first value is used as the first information;

About each predetermined project, the number that belongs to predetermined all the elements of organizing is counted, thereby obtained second value, and be worth second as second information; With

Use second value to make the first value normalization, thereby obtain the 3rd information.

Best, one group of predetermined content is one group of content broadcasting or distribute in predetermined a period of time.

Every group that constitutes by content as one of predetermined many groups content, described many group contents are broadcasting or many groups contents of distributing in different a period of times;

Calculate first value and second value of predetermined many groups content;

Use is corresponding to first value, and is the same as with first value second being worth that identical one group of content is calculated, and makes each first value normalization, thereby obtains the 3rd information.

Best, each predetermined many groups content is one group of content broadcasting or distribute in predetermined a period of time.

Best, the first information is relevant with content, discloses this content of indication and be the attribute information of the project of the content that the user likes.

Best, the first information is relevant with content, discloses this content of indication and be the attribute information of the project of the content that the user dislikes.

According to the present invention, for handling, thereby select the messaging device of the content that conforms to user's hobby that the 3rd information processing method is provided, described information processing method comprises:

First obtaining step of the first information of acquisition expression user's hobby;

Obtain second obtaining step of second information of the general hobby of expression; With

According to the first information that obtains in the processing of carrying out at first obtaining step and second information that in the processing that second obtaining step carries out, obtains, produce the hobby that discloses the user and produce step with respect to the deviation information of the 3rd information of the deviation of general hobby.

According to the present invention, provide preservation to carry out by computer, select the 3rd storage medium of computer program of the processing of the content that conforms to user's hobby, described computer program comprises:

According to the present invention, provide by computer and carry out, select the computer program of the processing of the content that conforms to user's hobby, described computer program comprises:

As mentioned above, according to the first information of the hobby of representative of consumer and second information of the general hobby of representative, produce the hobby that discloses the user the 3rd information with respect to the deviation of general hobby.

Description of drawings

Fig. 1 is the key diagram of the distribution of explanation television program broadcasting and stream data;

Fig. 2 is the block diagram of the structure of the Distributor shown in the presentation graphs 1;

Fig. 3 is that the explanation program vector produces the flow chart of processing 1;

Fig. 4 is the key diagram of explanation EPG data;

Fig. 5 is the key diagram of explanation program vector;

Fig. 6 is that the explanation program vector produces the flow chart of processing 2;

Fig. 7 is the flow chart of explanation packet transaction 1;

Fig. 8 is the flow chart of explanation packet transaction 2;

Fig. 9 is the flow chart of explanation title packet transaction 1;

Figure 10 is the flow chart of explanation title packet transaction 2;

Figure 11 is the flow chart of explanation title packet transaction 3;

Figure 12 is the flow chart of explanation title packet transaction 4;

Figure 13 is the block diagram of the structure of the program commending treatment facility shown in the presentation graphs 1;

Figure 14 is the flow chart that positive history vectors of explanation and negative history vectors produce processing 1;

Figure 15 is the key diagram of the positive history vectors of explanation;

Figure 16 is the flow chart that positive history vectors of explanation and negative history vectors produce processing 2;

Figure 17 is the flow chart of explanation matching treatment 1;

Figure 18 is the flow chart of explanation matching treatment 2;

Figure 19 is the flow chart of explanation matching treatment 3;

Figure 20 is the flow chart of explanation matching treatment 4;

Figure 21 is the flow chart of explanation matching treatment 5;

Figure 22 is that explanation user side's effect (effect) vector produces the flow chart of processing 1;

Figure 23 is that explanation user side effect vector produces the flow chart of processing 2;

Figure 24 is that explanation user side effect vector produces the flow chart of processing 3;

Figure 25 is that explanation user side effect vector produces the flow chart of processing 4;

Figure 26 is that explanation user side negative effect vector produces the flow chart of processing 1;

Figure 27 is that explanation user side negative effect vector produces the flow chart of processing 2;

Figure 28 is the flow chart that explanation comprises the matching treatment that group is recommended;

Figure 29 is the flow chart that explanation utilizes the matching treatment of user model;

Figure 30 is the flow chart of specification exception recommendation process;

Figure 31 is the block diagram of the structure of the TV receiving apparatus shown in the presentation graphs 1;

Figure 32 is the block diagram of the structure of the television display equipment shown in the presentation graphs 1;

Figure 33 is that the explanation recommendation information shows the flow chart of handling;

Figure 34 is the flow chart of explanation automatic channel set handling;

Figure 35 is the block diagram of the structure of the recording/reproducing apparatus shown in the presentation graphs 1;

Figure 36 is the flow chart of the automatic recording processing of explanation;

Figure 37 is the block diagram of another typical structure of expression Distributor;

Figure 38 is the block diagram of another typical structure of expression program commending treatment facility;

Figure 39 is the key diagram of another representative network of explanation distribution television program broadcasting and stream data;

Figure 40 is the block diagram of another typical structure of expression Distributor.

Embodiment

Below with reference to description of drawings one embodiment of the present of invention.

The distribution of television program broadcasting and stream data (streaming data) at first, is described with reference to figure 1.

Broadcasting station 1 is with the form of earthwave emission programming, perhaps 2 form emission programmings with the satellite ripple via satellite.Fig. 1 has only represented a broadcasting station 1.But needless to say, can there be a plurality of broadcasting stations.The antenna 3 that adopts in TV (TV) receiving equipment 4 receives the programming with the form emission of earthwave or satellite ripple.If desired, broadcast singal can comprise EPG (electronic program guides).

Distributor 5 reads stream data from stream data database 6, and by the network 8 that comprises internet and other subnet stream data is sent to TV receiving equipment 4.Distributor 5 is also read EPG or is read the metadata that comprises than the more detailed information of EPG from metadata database 7.EPG is and the 1 relevant information of broadcasting of program from the broadcasting station.Distributor 5 produces a program vector PP for each program subsequently, and by network 8, program vector PP is sent to EPG receiving equipment 9 together with the EPG data.

If the quantity that is included in the information among the EPG on the general broadcast signal that is superimposed on program is enough big for the processing that the back will illustrate, so be superimposed upon the general broadcast signal on the identical data of EPG can be used to described processing.On the other hand, if be included in not the processing of quantity of the information among the EPG that is superimposed on the general broadcast signal even as big as carrying out illustrating later, the EPG on being superimposed upon the general broadcast signal, metadata can be used to described processing so, perhaps is used as independent data.Owing to be included in not the processing of quantity of the information among the EPG that is superimposed on the general broadcast signal even as big as carrying out illustrating later, therefore in the present embodiment, comprise that the information of metadata is used to described processing, such information is called as the EPG data.

EPG receiving equipment 9 offers TV receiving equipment 4 to the EPG data that receive from Distributor 5.In addition, 9 program vector PP that receive together with the EPG data of EPG receiving equipment offer program commending treatment facility 10.

The control signal of the channel that the TV receiving equipment 4 with tuner is selected according to the indication user who receives from TV display device 11 with operation part or the recording/reproducing apparatus 12 that has operation part equally, select the broadcast singal of antenna 3, and receive the broadcast singal of selecting with the form reception of earthwave or satellite ripple.TV receiving equipment 4 also receives the stream data that transmits from Distributor 5 by network 8.In addition, TV receiving equipment 4 receives the EPG data from EPG receiving equipment 9, and these data are offered TV display device 11 or recording/reproducing apparatus 12.Be noted that if receive ripple to comprise EPG that TV receiving equipment 4 separates EPG with the programme signal that is included in the reception ripple so, and EPG and programme signal are offered TV display device 11 or recording/reproducing apparatus 12.

Program commending treatment facility 10 obtains program vector PP from EPG receiving equipment 9, and obtains Operation Log from TV display device 11 and recording/reproducing apparatus 12.Program commending treatment facility 10 is subsequently according to program vector PP and Operation Log, the perhaps input operation of carrying out according to the user, produce to recommend the recommendation information of the program that conforms to user's hobby, and recommendation information is offered TV display device 11 and recording/reproducing apparatus 12.

According to the operation input of user's input, 11 demonstrations of TV display device are from the broadcast singal of TV receiving equipment 4 receptions or the reproducing signal that receives from recording/reproducing apparatus 12.In addition, according to the recommendation information that receives from program commending treatment facility 10, TV display device 11 is provided with channel automatically, and shows about programs recommended information.TV display device 11 also offers program commending treatment facility 10 to the Operation Log of the history of the operation of carrying out as the user.

According to the operation input of user's input, recording/reproducing apparatus 12 is recorded in the recording medium of installation or the recording medium of embedding to the broadcast singal that receives from TV receiving equipment 4, for example on the hard disk.In addition, according to the recommendation information that receives from program commending treatment facility 10, recording/reproducing apparatus 12 records the broadcast singal that receives from TV receiving equipment 4 on the recording medium of the recording medium of installation or embedding automatically.On the other hand, recording/reproducing apparatus 12 is also from the recording medium of installation or the recording medium reproducing program of embedding, and a program that reproduces offers TV display device 11 so that show this program.In addition, recording/reproducing apparatus 12 also offers program commending treatment facility 10 to the Operation Log of the history of the operation of carrying out as the user.

In the superincumbent explanation, EPG receiving equipment 9, TV receiving equipment 4, program commending treatment facility 10, TV display device 11 is illustrated as different equipment with recording/reproducing apparatus 12.But these equipment needn't be designed separately.For example, needless to say, EPG receiving equipment 9, TV receiving equipment 4 and TV display device 11 can be integrated in the single structure of the TV receiver 15-1 with embedded tuner function.In addition, recording/reproducing apparatus 12 can be integrated with TV receiver 15-1, thereby form the TV receiver 15-2 with writing function.Needless to say, recording/reproducing apparatus 12 can be the so-called hdd recorder that comprises the hard disk with large storage capacity.In addition, program commending treatment facility 10 can be integrated with the TV receiver 15-1 with embedded tuner function, thereby form TV receiver 15-3, perhaps program commending treatment facility 10 can be integrated with the TV receiver 15-2 with writing function, thereby form TV receiver 15-4.

Fig. 2 is the block diagram of the structure of expression Distributor 5.

Data acquiring portion 21 obtains data from metadata database 7 or stream data database 6, and the data that obtain are offered tcp data segment 25.In addition, data acquiring portion 21 offers metadata to the EPG data and extracts part 22.In addition, data acquiring portion 21 is handled, thus according to the content of EPG data, be kept in the metadata database 7 the EPG data acquisition system in groups.

Metadata extracts part 22 and extract the necessary metadata of generation program vector PP from the EPG data that are received from data acquiring portion 21, and data that extract are offered program vector generation part 23.Program vector produces part 23 and produces program vector PP according to metadata, and if necessary, program vector PP be kept at before the vectorial EfPP of program side's effect (effect) in the storage part 24 offers tcp data segment 25, program vector PP is associated with program side effect vector EfPP.

If necessary, storage part 24 is used to preserve program side effect vector EfPP, and program side effect vector EfPP produces program vector PP information necessary data.

From the making side of finding out the measure that how to improve audience ratings and the viewpoint of broadcaster, general theatrical items are furnished with the characteristic of strongly-typed emphasized or content element, variety shows etc. are furnished with the characteristic of emphasizing strong performing artist's key element, and drama programs is furnished with the characteristic of emphasizing strong performing artist and playwright, screenwriter's key element.In order correctly to grasp the feature of program, must utilize these characteristics.That is, the type of program determines in the matching process that is carried out which component that constitutes the program vector PP of program is important, so that the program that recommendation conforms to user's hobby.In other words, such important component is different because of type.

That is, if the type of program is " general literature and art/documentary film ", so important project (item) is not the performing artist, but the content of program and title.On the other hand, if the type of program is " variety ", so important project is the performing artist.If the type of program is " drama ", so important project is performing artist and playwright, screenwriter.When described important project was used to produce program vector PP, the program effect vector was configured to be defined in every type the matching process, the percentage contribution of each project, and be stored in the storage part 24.

Tcp data segment 25 sends information to EPG receiving equipment 9 or TV receiving equipment 4 by network 8.EPG data and the stream data that is provided by data acquiring portion 21 is provided the information that transmits, and produces the program effect vector EfPP that part 23 provides by program vector.

If desired, driver 26 is connected with program vector generation part 23.If desired, disk 31, CD 32, magneto optical disk 33 or semiconductor memory 34 are installed on the driver 26, make data to exchange between driver 26 and disk 31, CD 32, magneto optical disk 33 or semiconductor memory 34.

Below with reference to the flow chart shown in Fig. 3, illustrate that the program vector that Distributor 5 carries out produces processing 1.

At first, at step S1, data acquiring portion 21 obtains to comprise the EPG data from the metadata of metadata database 7.

At step S2, metadata extracts the EPG data that part 22 receives from data acquiring portion 21 subsequently, and produces the required metadata of program vector PP from the EPG data pick-up.Subsequently, metadata extracts part 22 metadata that extracts is offered program vector generation part 23.

Fig. 4 has represented typical metadata.Metadata comprises as " the Movie:Japnese film " of the type of program with as " the Toukaidou Mitsuya Ghost-story " of the title of film.Metadata also comprises has broadcast the date, and publisher broadcasts the date, the title in broadcasting station and airtime.In addition, except the film of describing programme content explained orally, metadata also comprised director's name, playwright, screenwriter's name, cameraman's name, the personnel's of responsible music name and performing artist's name.

Subsequently, at step S3, if desired, program vector produces 23 pairs of information that are included in the metadata of part, and for example title and content are carried out morphology (morphological) analysis, and they are resolved into word.More particularly, program vector generation unit 23 resolves into 3 words to the movie name that is included in the metadata as title, i.e. " Toukaidou ", " Mitsuya " and " ghost-story ".As shown in Figure 4, metadata comprises the information of statement " 59 production ofSeihou; Masterpiece of horror show of Japanese film, depicting theworld of the famous Mitsuya ghost-story to fullness of formal beauty. ".In this case, the program vector generation part 23 extraction conducts information relevant with content is included in this and separates the word that is right.The word that extracts is " Seihou ", " formal ", " beauty ", " fullness ", " famous ", " Mitsuya ", " ghost-story ", " world ", " depicting ", " Japanese film ", " horror " and " Masterpiece ".

Subsequently, at next step S4, program vector produces part 23 project that is included in the metadata is converted to the vector that calls program vector PP.At last, the execution of end process.The program vector PP that produces is provided for tcp data segment 25, and tcp data segment 25 sends program vector PP to EPG receiving equipment 9 by network 8 subsequently.Program vector PP with described project can have the form that comprises all detail items components that are aligned to 1 array.As a kind of alternative, project is included in the large project, and described large project is converted into program vector PP subsequently.

Fig. 5 represented 7 sports, i.e. title, type, time period, broadcasting station, performing artist, playwright, screenwriter/author/producer and content application vector transformation and program vector PP={ Tm, Gm, Pm, Am, Km} of obtaining.The content of each sport is described as: title Tm={ title 1, and title 2 ..., type Gm={ drama, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other, time period (hour) Hm={ morning, daytime, night, prime time, the late into the night }, broadcasting station (TV station) Sm={NNK General, NNK Educational, AsianTelevision, TTS, Buji, Telenichi, Touto, First NNK Satellite, Second NNKSatellite, WOWO}, performing artist (personnel) Pm={A, B ..., playwright, screenwriter/author/producer Am={a, b ..., content (keyword) Km={kw1, kw2 ....

Because the type in these 7 sports, broadcasting station and time period all have the classification that can clearly discern, and the detail items of these each these sports can be represented by digital vectors.With the broadcasting station is example.As mentioned above, the broadcasting station is Sm={NNK General, NNKEducational, Asian Television, TTS, Buji, Telenichi, Touto, First NNKSatellite, Second NNK Satellite, WOWO}.For example, if the broadcasting station of program is WOWO, the broadcasting station vector can be by broadcasting station Sm={0 so, and 0,0,0,0,0,0,0,0,1} represents.With regard to type, sport is the Gm={ drama, variety, and physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other.If the type of program is general literature and art/documentary film, the type vector can be by type Gm={0 so, 0,0,0,0,0,1,0, and 0} represents.

On the other hand, the sport of title, performing artist, playwright, screenwriter/author/producer and content all can not be represented by digital vectors.In this case, each sport is by the vector representation that comprises as the frequency of occurrences of component of a vector.Be the frequency that this word occurs in sentence by the relevant frequency of occurrences of the word in the sentence of vector representation.For example, the sport of title be title=Toukaidou-1, Mitsuya-1, ghost-story-1}, wherein each component is a pair of word and represents the numeral of the frequency of this word.In this example, component Toukaidou-1 means that word Toukaidou occurs once in by the sentence of this vector representation.

Program vector PP is produced as mentioned above, is transmitted to EPG receiving equipment 9 subsequently.More particularly, the program vector PP that is produced by the program metadata of earlier in respect of figures 4 explanation is program vector PP={ title Tm={Toukaidou-1, Mitsuya-1, ghost-story-1}, type Gm={0,0,0,1,0,0,0,0,0}, time period Hm={0,0,0,0,1}, broadcasting station Sm={0,0,0,0,0,0,0,0,0,1}, performing artist Pm={Katsumi Wakasugi-1, ShigeruAmami-1, Toshihiko Emi-1, Ryuujirou Nakamura-1 and NorikoNishizawa-1}, playwright, screenwriter/author/producer Am={Nobuo Nakagawa-1, ShochiOhnuki-1, Yoshihiro Ishida-1, Shoji Nishimoto-1, Chuta Watanabe-1}, content Km={Seihou-1, formal-1, beauty-1, fullness-1, famous-1, Mitsuya-1, ghost-story-1, world-1, depicting-1, Japanese-1, horror-1 and masterpiece-1}}.

Program vector PP is transmitted to EPG receiving equipment 9.

In the program vector PP that produces as mentioned above, the fundamental component of performing artist's vector Pm can be by special weighting.For the same reason, the director of playwright, screenwriter/author/producer's vector Am can be by special weighting.For example, performing artist's vector can be by Pm={KatsumiWakasugi-3, Shigeru Amami-2, Toshihiko Emi-1, RyuujirouNakamura-1 and Noriko Nishizawa-1} represent that playwright, screenwriter/author/producer's vector can be by Am={Nobuo Nakagawa-3, Shochi Ohnuki-1, Yoshihiro Ishida-1, ShojiNishimoto-1, Chuta Watanabe-1} represents.

In addition, as mentioned above, the type of program determines that in order to recommend the program that conforms to user's hobby in the matching process that carries out, which is important in the component of the program vector PP of formation program, and in other words, such important component is different because of type.Program side effect vector EfPP shows which component is important concerning every type.If program side effect vector EfPP is stored in the storage part 24, so also can transmit program side effect vector EfPP by being associated with program vector PP.One group of sport by the program vector PP representative is provided with program side effect vector EfPP.

The hypothetical program vector PP=and title Tm, type Gm, time period Hm, broadcasting station Sm, performing artist Pm, playwright, screenwriter/author/producer Am, the type of the program among the content Km} is general literature and art/documentary film.In this case, important project is title and content.Thereby, the program side effect vector EfPP value of being configured to 3,1,1,1,1,1, and 3}, wherein numeral 3 is weighted values.On the other hand, if the type of program is a variety, so important project is the performing artist.In this case, the program side effect vector EfPP value of being configured to 1,1,1,1,5,1, and 1}, wherein numeral 5 is weighted values.If the type of program is a drama, so important project is performing artist and playwright, screenwriter.In this case, the program side effect vector EfPP value of being configured to 1,1,1,1,2,3, and 1}, wherein numeral 2 is the weighted values that are used for the performing artist, numeral 3 is the weighted values that are used to write a play.

By with reference to the flow chart shown in the figure 6, following interpretation is wherein by being associated with program vector PP, and the program vector that transmits the situation of program side effect vector EfPP produces handles 2.

The processing of carrying out at step S11-S14 is identical with the processing of carrying out at the step S1-S4 with reference to the flow chart of figure 3 explanations respectively.In these steps, as previously mentioned, obtain the EPG data from metadata database 7, and produce the necessary metadata of program vector PP from the EPG data pick-up.Subsequently, if desired, the information that is included in the metadata is carried out morphological analysis, so that information decomposition is become word.The example that is included in the information in the metadata is title and content.Subsequently, the project that is included in the metadata is converted into program vector PP.

Subsequently, at next step S15, program vector produces the type of part 23 according to the program relevant with the metadata that receives, extraction effect vector in the effect vector information from be kept at storage part 24.If the type of program is a variety, extract the effect vector EfPP={1 that the performing artist is provided with

weighted value

5,1,1,1,5,1,1} from storage part 24 so.

Subsequently, at next step S16, program vector produces part 23 is associated the program vector PP that produces among the extracted effect vector EfPP and the processing of carrying out at step S14 in the processing that step S15 carries out.At last, the execution of end process.

By carrying out above-mentioned processing, program vector PP is produced, and is associated with program side effect vector EfPP that the big event to program vector PP that obtains according to the type of program is provided with weighted value.Subsequently, by network 8, program vector PP and program side effect vector EfPP are transmitted to EPG receiving equipment 9.

Program vector PP produces as mentioned above.But, program is divided into groups available less processing, but the higher real estate tight knot of accuracy order vector PP by attribute according to program.

General in order to produce by the program with same type, the group that the serial of for example broadcasting in same week is formed perhaps by arranging the unit at program, is for example introduced the group of the identical table person's of drilling program composition, the processing of the program that divides into groups in 13 weeks.As the object lesson of processing of grouping program, by flow chart with reference to figure 7, the packet transaction 1 of the grouping program of following interpretation serial.

At first, at step S31, extract the program that conforms to predetermined branch set condition in the EPG data of data capture unit 21 from be kept at metadata database 7.As an example of predetermined packet of packets condition, the program that this conditional request will be extracted has identical title, and identical broadcasting station is in identical broadcasts time of all working day or in identical broadcasts time on identical date in each week.

Subsequently, at next step S32, data capture unit 21 is put into group to the program that extracts, and to the additional group of the EPG of program data ID.

Subsequently, at next step S33, metadata extracts part 22 and extracts in order to produce the program that representative is identified as the part of serial, i.e. first program or by the program vector PP of first broadcasting of the program of ID identification and the metadata that needs on the same group mutually.

Subsequently, at next step S34, the program vector of carrying out the flowchart text shown in earlier in respect of figures 3 or 6 produces to be handled.

Subsequently, at next step S35, program vector produces the program vector PP that part 23 is fixed the program vector PP of the generation of first broadcasting and is arranged to be discerned by group ID.At last, the execution of end process.

By carrying out above-mentioned processing, the program of serial is placed in the group with same program vector PP.In addition, can make the ID of group and program vector PP interrelated, and be kept in the storage part 24.

Especially, in the EPG of serial data, in many cases, first content is the description of all programs, and subsequent content all is the description of related-program, rather than the description of all programs.In addition, for each program, the metadata part except program is described is identical.Thereby, produce program vector PP by utilizing first segment purpose EPG data, not only can reduce and carry out the number of times that program vector produces processing, but also can produce the program vector PP that pin-point accuracy ground conforms to the characteristic of program.

Fig. 7 is that explanation is the flow chart of the processing in the EPG data of public ID adding serial.If the EPG data have comprised the information of discerning serial, the processing of carrying out at step S31 can be eliminated so.In this case, by discerning the information of serial with reference to being included in being used in the EPG data, public ID is added in the EPG data of serial.

By with reference to figure 8, following explanation is also about being different from the element of serial, the packet transaction 2 that explanation can be divided into groups to program.Based on shown in being grouped as follows of type and performing artist.

At first, at step S51, data acquiring portion 21 is with reference to the EPG data that are kept in the metadata database 7, and the cluster code corresponding to the metadata of program is joined in the EPG data of program.

For example suppose that the program that conforms to the condition of serial is extracted.In general, the program that the conditional request of serial will be extracted has same title, and the identical broadcasts radio station is in identical broadcasts time of all working day or in identical broadcasts time on identical date in each week.Extract program for one group of the program that is identified as serial, have the least significant digit of determining according to the airtime of serial general as described below corresponding to first cluster code of serial.If the airtime of the program of formation serial is the identical time in all working day, the least significant digit of first cluster code is the code for 1 so.On the other hand, be in the identical broadcasts time on the identical date in each week if constitute the airtime of the program of serial, the least significant digit of first cluster code is the code for 2 so.If constituting the airtime of the program of serial is except that corresponding to the time the time of

code

1 and 2, for example on week that replaces or date, the least significant digit of first cluster code is for example to be 3 code so.For the program except that serial, the least significant digit of first cluster code is for example to be 0 code.

Subsequently, the categorical data in the metadata of reference record in metadata database 7 is so that determine corresponding to second numerical digit in the least significant digit in second cluster code of type.According to type, it for example is 20 code that second numerical digit in the least significant digit in second cluster code is configured to, and is code of 30 or the like.Second numerical digit in the least significant digit in second cluster code is so-called ten.

At last, the performing artist's data in the metadata of reference record in metadata database 7, so that determine corresponding to the 3rd numerical digit in the least significant digit in performing artist's the 3rd cluster code, and the position preface is higher than other numerical digit of the 3rd numerical digit.According to the performing artist, it for example is 2300 code that other numerical digit that the 3rd numerical digit in the least significant digit in the 3rd cluster code and position preface are higher than the 3rd numerical digit is configured to, and is 800 code or other code.The 3rd numerical digit in the least significant digit in the 3rd cluster code is so-called hundred, other numerical digit that the position preface is higher than the 3rd numerical digit from but a position preface is higher than other numerical digit of hundred.

Cluster code among the EPG of adding program is the summation of above-described first to the 3rd cluster code.

Thereby at next step S52, whether data acquiring portion 21 is distributed to the cluster code of EPG by inspection first numerical digit is 0, determines whether EPG is the EPG of serial.

If it is the program of serial that the definite result who produces in the processing that step S52 carries out indicates program, handling process enters step S53 and S54 so, carries out respectively the step S33 processing identical with S34 with the flow chart of earlier in respect of figures 7 explanations.That is, the EPG of first broadcasting is extracted, and carries out the program vector generation processing of the flowchart text shown in earlier in respect of figures 3 or 6.

Subsequently, at next step S55, program vector produces part 23 program vector PP of first broadcasting is fixed as the program vector PP of serial, and the cluster code of generation is associated with program vector PP.At last, the execution of end process.

On the other hand, not the program of serial if the definite result who produces in the processing that step S52 carries out indicates program, handling process enters step S56 so, and the program vector of carrying out the flowchart text shown in the earlier in respect of figures 3 or 6 produces to be handled.

Subsequently, at next step S57, program vector produces part 23 makes the cluster code of generation be associated with program vector PP.At last, the execution of end process.

By carrying out above-mentioned processing, not only can divide into groups to program vector PP, and can divide into groups to program vector PP according to type and performing artist according to serial.In addition, the cluster code of identification group is associated with the program vector PP of this group.

In front in the Distributor 5 with reference to figure 2 explanations, as above with reference to the described such program vector PP that produces of figure 3-8.Thereby, always can be with corresponding to new terminology, the form of the vector of newtype etc. produces program vector PP.The program vector PP and the EPG data thereof that produce are transmitted to the EPG receiving equipment by network 8, and are provided for program commending treatment facility 10.

In addition, packet transaction can comprise the morphological analysis of carrying out to title, so that title is resolved into word, and to a group of word distribution ID.

Below with reference to the flowchart text title packet transaction 1 shown in Fig. 9.

At first, at step S61, data acquiring portion 21 is from by with reference to being kept at extracting header in the catalogue metadata that the EPG data the metadata database 7 obtain, and title is submitted to program vector produces part 23.

Subsequently, at next step S62, program vector produces 23 pairs of titles of part and carries out morphological analysis, and title is resolved into word.More particularly, suppose that the movie title that is included in the metadata is " Toukaidou Mitsuya Ghost-story ".In this case, this title is broken down into following 3 word and: " Toukaidou ", " Mitsuya " and " Ghost-story ".

Subsequently, at next step S63, program vector produces part 23 and extract a word (phrase that perhaps comprises a plurality of words) from the word (or phrase) that obtains as the result who analyzes, and from 24 extractions of storage part the group ID of the word that extracts (or phrase).

The phrase that comprises a plurality of words is one group of word.Produce this group word by combination as the word that the result of syntactic analysis obtains.Suppose that 3 word and that obtain as the result of morphological analysis are " Toukaidou ", " Mitsuya " and " Ghost-story ".In this case, this group word can be " Toukaidou Mitsuya ", " Toukaidou Ghost-story " or " Mitsuya Ghost-story ".

Subsequently, at next step S64, program vector produces part 23 and determines whether corresponding group ID is extracted from storage part 24.

If the definite result's indication that produces in the processing that step S64 carries out is not extracted corresponding group ID from storage part 24, handling process enters step S65 so, and program vector produces part 23 makes new group ID be associated with the word of extraction (extracting phrase/group of words that perhaps comprises a plurality of words).This is because there is not corresponding group ID to be stored in the storage part 24, and is associated with the word (extracting phrase/group of words that perhaps comprises a plurality of words) that extracts.Subsequently, program vector generation part 23 is kept at this word (extracting phrase/group of words that perhaps comprises a plurality of words) in the storage part 24 together with relevant group ID.

On the other hand, if the definite result's indication that produces in the processing that step S64 carries out has been extracted corresponding group ID from storage part 24, handling process enters step S66 so.Finish after the processing that step S65 carries out, handling process also enters step S66.At step S66, program vector produces part 23 and determines whether each word (perhaps each group of words) that constitutes title have been extracted a group ID.

If the definite result's indication that produces in the processing that step S66 carries out is not also extracted a group ID to each word (perhaps each group of words) that constitutes title, handling process is returned step S63 so, once more the processing of execution in step S63 and subsequent step.

On the other hand, if the definite result's indication that produces in the processing that step S66 carries out has been extracted a group ID for each word (perhaps each group of words) that constitutes title, handling process enters step S67 so, and program vector produces part 23 makes the group ID of extraction be associated with program vector PP.At last, the execution of end process.

By carrying out above-mentioned processing, the group ID of the word (perhaps group of words) that constitutes title is associated with program vector PP corresponding to this title.Tcp data segment 25 sends group ID and title to TV receiving equipment 4 or EPG receiving equipment 9 by network 8 subsequently.

In addition, in the program with similar title can be included in mutually on the same group.For example, by the program title word that relatively obtains as the result of the morphological analysis that the word that constitutes program title is carried out, calculating is in the predetermined cycle, the similarity between the program title that occurs in 2 weeks, 1 month or half a year for example, title be the serial of " Ginpachi; Second-Grade A Class Teacher " can be placed into title in the identical group of the special series of " Ginpachi Special, Second-Grade A ClassTeacher ".Have only when the similarity of calculating equals predetermined value at least, serial just is placed in the identical group with special series.

By the flow chart shown in reference Figure 10, following interpretation is according to the word that constitutes one of title and constitute matching degree between the word of another title, to the title packet transaction 2 of title grouping.

The processing of carrying out at step S401 and S402 is with identical with reference to the step S61 of the flow chart of figure 9 explanations and processing that S62 carries out in front respectively.That is, data acquiring portion 21 is from by with reference to being kept at extracting header in the catalogue metadata that the EPG data the metadata database 7 obtain, and title is submitted to program vector produces part 23.Subsequently, program vector produces part 23 each title is carried out morphological analysis, and each title is resolved into word.

Subsequently, at next step S403, program vector produces part 23 calculating from the matching degree between the word of the title of morphological analysis acquisition.That is, program vector produces the matching rate that part 23 is found out the word of another title of word matched of representing one of title.

More particularly, we are example with title " Ginpachi, Second-Grade A ClassTeacher " and title " Ginpachi Special, Second-Grade A Class Teacher ".Title " Ginpachi, Second-Grade A Class Teacher " is broken down into following word: " Ginpachi ", " Second ", " Grade ", " A ", " Class " and " Teacher ".On the other hand, title " Ginpachi Special, Second-Grade A Class Teacher " is broken down into following word: " Ginpachi ", " Special " " Second ", " Grade ", " A ", " Class " and " Teacher ".Subsequently, to word " Ginpachi ", " Second ", " Grade ", " A ", " Class " and " Teacher " and word " Ginpachi ", " Special " " Second ", " Grade ", " A ", " Class " and " Teacher " carries out morphological analysis, thereby is produced as 6/7 or 85.7% matching rate.

Subsequently, at next step S404, program vector produces part 23 and determines whether the matching rate of words equals predetermined value at least, and for example 70%.Needless to say, the threshold value of matching rate can be to be different from 70% value.

If the definite result who produces in the processing that step S404 carries out indicates the matching rate of word to equal predetermined value at least, for example 70%, handling process enters step S405 so, and program vector produces part 23 makes identical group ID be associated with these programs.Subsequently, program vector produces part 23 couplings word or coupling group of words and organizes ID and is kept at together in the storage part 24.

On the other hand, if the matching rate that the definite result who produces in the processing that step S404 carries out indicates word less than predetermined value, for example 70%, handling process enters step S406 so.After the processing that step S405 carries out finished, handling process also entered step S406.At step S406, whether program vector produces all words that part 23 determines one of titles and compares with all words of another title.

If the definite result who produces in the processing that step S406 carries out indicates all words of one of title also not have to compare with all words of another title, handling process is returned step S403 so, carries out the processing of step S403 and the processing of subsequent step once more.

On the other hand, if all words that the definite result who produces in the processing that step S406 carries out indicates one of title compare the execution of end process so with all words of another title.

By carrying out above-mentioned processing, the group ID based on the matching rate of the word that constitutes title is associated with the program vector PP of the program with this title.Subsequently, tcp data segment 25 sends group ID and title to TV receiving equipment 4 or EPG receiving equipment 9 by network 8.Thereby, during the program with similar title can be included into mutually on the same group.Example with program of similar title is aforesaid serial and special series.

In addition, by matching rate generation group according to the word that constitutes title, program with same title can be detected as same group program, although EPG or metadata comprise by half-size scale and full-scale numeral, the symbol difference in appearance that half-size scale and full-scale character or lowercase character and upper case character cause.

In addition, except the matching rate of word, the broadcasting station, program category or broadcasting time started can be added in the branch set condition.We are example with the news program with title " News ".Because this title only has the seldom word that comprises " News ", therefore in the processing of explaining with reference to the flow chart shown in Figure 10 in the above, different broadcasting stations and multi-form news program may be detected as the program that belongs to mutually on the same group.Thereby, so that address this problem, just have only during the program of the matching rate that satisfies word of identical broadcasts station broadcast is included into mutually on the same group.

By the flow chart shown in reference Figure 11, below interpretation except according to the word that constitutes one of title with constitute the matching degree between the word of another title, also as requested the additional conditions in identical broadcasts radio station to the title packet transaction 3 of title grouping.

The processing of carrying out to S424 at step S421 is with the step S401 with reference to the flow chart of Figure 10 explanation is identical to the processing that S404 carries out in front respectively.That is, data acquiring portion 21 is from by with reference to being kept at extracting header in the catalogue metadata that the EPG data the metadata database 7 obtain, and title is submitted to program vector produces part 23.Subsequently, program vector produces part 23 each title is carried out morphological analysis, and each title is resolved into word.Subsequently, program vector produces part 23 calculating from the matching degree between the word of the title of morphological analysis acquisition.That is, program vector produces the matching rate that part 23 is found out the word of another title of word matched of representing one of title.Subsequently, program vector produces part 23 and determines whether the matching rate of words equals predetermined value at least, and for example 70%.

If the definite result who produces in the processing that step S424 carries out indicates the matching rate of word to equal predetermined value at least, for example 70%, handling process enters step S425 so, and program vector produces part 23 and whether determines to have the program of described title by identical broadcasting station broadcasting.

If the definite result who produces in the processing that step S425 carries out indication has the program of described title by identical broadcasting station broadcasting, handling process is carried out step S426 so, and program vector produces part 23 makes program be associated with identical group ID.Subsequently, program vector generation part 23 is kept in the storage part 24 together mating word or coupling group of words and group ID and broadcasting station.

On the other hand, if the definite result who produces in the processing that step S424 carries out indicates the matching rate of word less than predetermined value, for example 70%, the program that the definite result's indication that perhaps produces in the processing that step S425 carries out has described title is not that handling process enters step S427 so by identical broadcasting station broadcasting.After the processing that step S426 carries out finished, handling process also entered step S427.At step S427, whether program vector produces all words that part 23 determines one of titles and compares with all words of another title.

If the definite result who produces in the processing that step S427 carries out indicates all words of one of title also not have to compare with all words of another title, handling process is returned step S423 so, carries out the processing of step S423 and the processing of subsequent step once more.

On the other hand, if all words that the definite result who produces in the processing that step S427 carries out indicates one of title compare the execution of end process so with all words of another title.

By carrying out above-mentioned processing, make matching rate, and be associated with the program vector of program with described title based on the group ID of the program with described title by the fact of identical broadcasting station broadcasting based on the word that constitutes title.Subsequently, tcp data segment 25 sends group ID and title to TV receiving equipment 4 or EPG receiving equipment 9 by network 8.Thereby, in the processing of the program with similar title being included into mutually on the same group, prevent that program by the broadcasting of different broadcasting stations is included into mutually on the same group.Have similar title, but be aforesaid news program by the example of the program of different broadcasting stations broadcasting.

In the above in the processing with reference to the flowchart text shown in Figure 11, except the matching rate of the word that requires to constitute title should equal the condition of predetermined value at least, it should be program by identical broadcasting station broadcasting that additional conditions require program.But, be noted that except the broadcasting station airtime section and type also can be included as the condition to the title grouping of program the condition that matching rate except the word that requires to constitute title should equal predetermined value at least certainly.

In addition, even because the cause of real sports report or special series, the broadcasting time started of serial or the program play in the set time is delayed, at least should equal on the same group the condition of program being included into mutually the condition of predetermined value as matching rate except the word that requires to constitute title, whether the broadcasting time started by determining program is in the preset time scope, for example in 1 hour, can divide into groups to program.

By the flow chart shown in reference Figure 12, below interpretation except according to the word of one of title that constitutes program with constitute the matching degree between the word of another title, also according to about broadcasting time started of program definite result in the scope whether at the fixed time, to the title packet transaction 4 of the title grouping of program.

The processing of carrying out to S444 at step S441 is with the step S401 with reference to the flow chart of Figure 10 explanation is identical to the processing that S404 carries out in front respectively.That is, data acquiring portion 21 is from by with reference to being kept at extracting header in the catalogue metadata that the EPG data the metadata database 7 obtain, and title is submitted to program vector produces part 23.Subsequently, program vector produces part 23 each title is carried out morphological analysis, and each title is resolved into word.Subsequently, program vector produces part 23 calculating from the matching degree between the word of the title of morphological analysis acquisition.That is, program vector produces the matching rate that part 23 is found out the word of another title of word matched of representing one of title.Subsequently, program vector produces part 23 and determines whether the matching rate of words equals predetermined value at least, and for example 70%.

If the definite result who produces in the processing that step S444 carries out indicates the matching rate of word to equal predetermined value at least, for example 70%, handling process enters step S445 so, whether program vector produces broadcasting time started of program that part 23 determines to have described title in the preset time scope, is for example moved mutually in 1 hour.

If the broadcasting time started that the definite result's indication that produces in the processing that step S445 carries out has the program of described title is moved in the preset time scope mutually, handling process is carried out step S446 so, and program vector produces part 23 makes program be associated with identical group ID.Subsequently, program vector produces part 23 couplings word or coupling group of words and group ID and broadcasts the scope of time started and is kept at together in the storage part 24.

On the other hand, if the definite result who produces in the processing that step S444 carries out indicates the matching rate of word less than predetermined value, for example 70%, the broadcasting time started that the definite result's indication that perhaps produces in the processing that step S445 carries out has the program of described title is not moved in the preset time scope mutually, and handling process enters step S447 so.After the processing that step S446 carries out finished, handling process also entered step S447.At step S447, whether program vector produces all words that part 23 determines one of titles and compares with all words of another title.

If the definite result who produces in the processing that step S447 carries out indicates all words of one of title also not have to compare with all words of another title, handling process is returned step S443 so, carries out the processing of step S443 and the processing of subsequent step once more.

On the other hand, if all words that the definite result who produces in the processing that step S447 carries out indicates one of title compare the execution of end process so with all words of another title.

By carrying out above-mentioned processing, make matching rate, and in the preset time scope, be associated with the program vector of program based on the broadcasting time started of program with described title by the group ID of the fact that moves mutually with described title based on the word that constitutes title.Subsequently, tcp data segment 25 sends group ID and title to TV receiving equipment 4 or EPG receiving equipment 9 by network 8.Thereby, except according to the word of one of title that constitutes program with constitute the matching degree between the word of another title, also according to about broadcasting time started of program definite result in the preset time scope whether, program with similar title is included in the phase processing on the same group, and the broadcasting time started is detected as the program that belongs to identical by the program that moves mutually in the preset time scope.Thereby can prevent that in the preset time scope broadcasting time started is not detected as by mobile mutually programs such as special serieses and belongs to phase program on the same group.

Figure 13 is the block diagram of the structure of expression program commending treatment facility 10.

Data acquiring portion 41 is from Distributor 5 program receiving vector PP, and and the assembly of the program side effect vector EfPP that transmits together of program vector PP.It is necessary from the program vector PP extraction matching treatment of Distributor 5 receptions that program vector extracts part 42, perhaps produces the assembly of the necessary program vector PP of user model.If desired, program vector extracts part 42 program that extracts is sent to matching treatment part 43 together with the program side effect vector EfPP corresponding to the program vector PP that extracts.

Operation input section 44 generally comprises the input unit such as keyboard, touch pads and mouse.Operation input section 44 is to receive the initial inventory information of user's input and the input theme that is used to produce user model, and described information and input theme is offered the assembly of initial directory stores part 45.Initial directory stores part 45 is to the initial inventory information that receives from operation input section 44 and produces the assembly of the theme catalogue of user model equally from operation input section 44 being used to of receiving.If desired, 45 initial inventory informations of initial directory stores part and theme offer Operation Log and obtain part 46 or matching treatment part 43.According to the operation of user, upgrade the data that are kept in the initial directory stores part 45 frequently by operation input section 44 inputs.Initial inventory information generally comprises the information that discloses the program that the user dislikes, and the information that discloses the program that the user likes.The example of the information of the program that the announcement user dislikes is the type of disliking, keyword of disliking and the performing artist who dislikes.On the other hand, the example of the information of the program that the announcement user prefers is the type of liking, keyword of liking and the performing artist who likes.

It is to obtain Operation Log from TV display device 11 or recording/reproducing apparatus 12 that Operation Log obtains part 46, and the information of Operation Log is divided into positive history and negative historical assembly.If desired, Operation Log obtains part 46 with reference to the information that is kept in the initial directory stores part 45, the program vector PP that obtains from data acquiring portion 41 extracts the program vector PP corresponding to positive and negative history, and the positive and negative history corresponding to the program vector PP of extraction is offered positive historical storage part 47 and negative historical storage part 48 respectively.Positive historical storage part 47 is preserved the incident of the positive history of supplying with it, and produces positive history vectors UP.By the same token, negative historical storage part 48 is preserved the incident of the negative history of supplying with it, and produces negative history vectors MUP.The positive history vectors UP that produces and the negative history vectors MUP of generation are provided for matching treatment part 43.

Above-cited positive history is to be used to extract the user to want the program watched in a hurry, perhaps in other words, and the information of a program that is considered to miss potter.When the user watches and during recorded program, perhaps more particularly, when the user is received in the processing that illustrates later, the program of from programs recommended tabulation, advising to the user, and watch and when writing down the program of acceptance, the metadata of this program is stored in the positive historical storage part 47 as impressive metadata.Positive historical storage part 47 obtains the summation of the positive history of each detail items or each sport, thereby produces positive history vectors UP.

On the other hand, above-cited negative history is the information that is used for from the undesirable program of programs recommended eliminating.Undesirable program is that the user wants the program watched hardly.When the user did not watch and writes down undesirable program, the metadata of this program was stored in the negative historical storage part 48 as unimpressive metadata.The example of undesirable program is the program with the project correspondence of stipulating in the information of initial record of disliking, not viewed after the record is deleted program, in the back in Shuo Ming the processing, to user's suggestion from programs recommended tabulation, but the program of not accepted by the user.Suppose that the user is defined as the project of disliking to physical culture in the information of initial record.In this case, comprise the type Gmup={0 that is used for as the weighted value 5 of the physical culture of bearing impression, 0,5,0,0,0,0,0,0} is stored in the negative historical storage part 48 as additional information.For the same reason, negative historical storage part 48 obtains the summation of the negative history of each detail items or each sport, thereby produces negative history vectors MUP.

Matching treatment part 43 checking from program vector extract program vector PP that part 42 extracts and the positive history vectors UP that receives from positive historical storage part 47 between, the perhaps coupling between program vector PP and the negative history vectors MUP that receives from negative historical storage part 48.

Program vector PP, positive history vectors UP and negative history vectors MUP can be the vectors with the form that comprises all detail items components that are aligned to 1 array.In this case, because the information such as title or keyword has a plurality of words, so in vector, the weighted value that each word has and the project such as type equates.In order to address this is that, the normalized part 61 that adopts in matching treatment part 43 makes title and the word normalization that constitutes described title as the project that is made of word by the number of the frequency of each word divided by the word of the title that constitutes each program.We are with title Tm={Toukaidou-1, and Mitsuya-1, ghost-story-1} are example.In this case, this title is normalized to Tm={Toukaidou:0.33, Mitsuya:0.33, ghost-story:0.33}.Like this, for as the title of the project that is made of word with constitute the word of described title, the summation of normalized frequency that constitutes the word of described title is used as weighted value, and because summation is 1, therefore any problem can not occur in matching process.

Vector operation part 62 is used for determining between program vector PP and the positive history vectors UP, and the assembly of the matching treatment of the coupling between program vector PP and the negative history vectors MUP.

If program vector PP, positive history vectors UP and negative history vectors MUP are the vectors with the form that comprises all detail items components that are aligned to 1 array, vector operation part so 62 obtains the similarity SimUP between program vector PP and the positive history vectors UP, and the similarity SimMUP between program vector PP and the negative history vectors MUP.Similarity SimUP and similarity SimMUP use cosine distance c os θ to represent.Calculate the cos θ u that represents similarity SimUP according to equation given below (1), and calculate the cos θ m that represents similarity SimMUP according to the equation that provides equally below (2).Can find out from equation (1) and (2),, calculate the cosine distance by the product of the inner product of two vectors divided by these two absolutes value of a vector.

SimUP＝cosθu＝UP·PP/|UP|×|PP| (1)

SimMUP＝cosθm＝MUP·PP/|MUP|×|PP| (2)

In superincumbent equation (1) and (2), symbol PP, UP and MUP represent program vector PP respectively, positive history vectors UP and negative history vectors MUP.In addition, operator of symbol " " expression is used to obtain the inner product at the vector of these operator both sides.On the other hand, operator of symbol " * " expression is used to obtain the scalar product at the absolute value of a vector of these operator both sides.

In addition, if program vector PP, positive history vectors UP and negative history vectors MUP are the vectors that obtains as the result of the vectorization processing that each sport is carried out, vector operation part so 62 can obtain for each sport, similarity between program vector PP and the positive history vectors UP, and the similarity between program vector PP and the negative history vectors MUP, and the summation of the similarity that obtains about all sports is calculated as each similarity SimUP and similarity SimMUP.

For example suppose for being broken down into word, and be used as the title of sport, the positive history vectors UP that provides represents title Tup={school-1, ghost-story-1, toilet-1}, and the program vector PP that provides is represented title Tm={Toukaidou-1, Mitsuya-1, ghost-story-1}.The length of our phantom order bit vector is 1 in addition.In this case, the length (absolute value) of vector is square root sum square of component, calculates the sine and cosine distance of representing two similarities between the title according to following equation (3):

cosθt＝1·1/×＝ (3)

In equation (3), operator of symbol " " expression is used to obtain the inner product at the vector of these operator both sides, and operator of symbol " * " expression, is used to obtain the scalar product at the absolute value of a vector of these operator both sides.

For the same reason, according to the mode that is similar to equation (3), for the title that is used separately as sport, the negative cosine distance of the similarity between two titles that can obtain representing to represent by program vector PP and negative history vectors MUP.

Now, for example suppose that positive history vectors UP is compound positive history vectors UP={ title Tup, type Gup, performing artist Pup, playwright, screenwriter/author/producer Aup, content (keyword) Kup}, negative history vectors MUP is compound negative history vectors MUP={ title Tmup, type Gmup, performing artist Pmup, playwright, screenwriter/author/producer Amup, content (keyword) Kmup}.In this case, for title as sport, type, the performing artist, each in playwright, screenwriter/author/producer and the content is able to sine and cosine distance and negative cosine distance according to equation (3) according to identical mode.Owing to each sport is obtained sine and cosine distance and negative cosine distance, therefore all calculates similarity SimUP and similarity SimMUP according to equation as follows (4):

Sim＝cosθ _t+cosθ _g+cosθ _p+cosθ _a+cosθ _k (4)

In the superincumbent equation, cos θ _tBe about the title sport, between program vector PP and the positive history vectors UP, the perhaps cosine distance between program vector PP and the negative history vectors MUP.Cos θ _gBe about the type sport, between program vector PP and the positive history vectors UP, the perhaps cosine distance between program vector PP and the negative history vectors MUP.Cos θ _pBe about performing artist's sport, between program vector PP and the positive history vectors UP, the perhaps cosine distance between program vector PP and the negative history vectors MUP.Cos θ _aBe about playwright, screenwriter/author/producer's sport, between program vector PP and the positive history vectors UP, the perhaps cosine distance between program vector PP and the negative history vectors MUP.Cos θ _kBe about content (keyword) sport, between program vector PP and the positive history vectors UP, the perhaps cosine distance between program vector PP and the negative history vectors MUP.

If as shown in top equation, similarity SimUP and similarity SimMUP are calculated as the summation of the cosine distance that obtains about all sports, and the deviation of the weighted value aspect between the project is eliminated so.Thereby, the principle that top calculating employing is identical with the normalization that illustrates previously.Thereby, being different from program vector PP, positive history vectors UP and negative history vectors MUP have the form that comprises all detail items components that are arranged in 1 array, and the normalized of being undertaken by normalized part 61 can be omitted.

Promptly, if under the situation of not carrying out normalized, use and all have the program vector PP of the form that comprises all detail items components that are arranged in 1 array, positive history vectors UP and negative history vectors MUP, so in the project such as title and content, increase along with the number of historical events, the number of different words increases, but with have the project that is easy to overlapping component and compare, the frequency of each various words increases hardly.Example with the project that is easy to overlapping component is broadcasting station and type.

For this reason, if similarity SimUP and similarity SimMUP all be calculated as about the summation of the important cosine distance that obtains, the effect that so all has each project that is easy to overlapping component becomes bigger.Example with the project that is easy to overlapping component is aforesaid broadcasting station and type project.We suppose that the user is the admirer of commentator A, thereby the user likes watching " commentary of the baseball game of the B of team that is explained orally by commentator A ".In this case, the information " commentary of baseball game " as the type project is easy to overlapping for historical.On the other hand, the information " commentator A " as performing artist's project is difficult to overlapping for historical.Thereby in some cases, the commentary of the baseball game of the B of team that another commentator B explains orally is recommended, rather than recommends to introduce the variety show of commentator A.

If carry out normalized, perhaps similarity SimUP and similarity SimMUP are calculated as the summation of the cosine distance that obtains about all sports, and the variety show of commentator A participation can be recommended so, and not influenced by the value of historical frequency.Thereby the hobby that can pin-point accuracy ground in recommend programs, reflects the user.

In addition, vector operation part 62 can also basis and the program side effect vector EfPP that transmits together of program vector PP and the information that is kept at the initial catalogue of user in the initial directory stores part 45, perhaps according to the generation that illustrates later and in user profile catalogue part 63 user side's effect vector EfUP of catalogue or the user side's negative effect vector EfMUP that illustrates later use weighted value, calculate similarity SimUP and similarity SimMUP.

Subsequently, according to similarity SimUP by adopting said method to calculate, vector operation part 62 is also from having the program of highest similarity, calculate the similarity SimMUP to negative history vectors MUP of the program of predetermined number, the program of described predetermined number all has the higher similarity SimUP that aligns history vectors UP.In general, the predetermined number of such program is set to 10.Subsequently, vector operation part 62 obtains difference (SimUP-SimMUP), and from having the program of maximum difference (SimUP-SimMUP), determines all to have the program than the predetermined number of big difference (SimUP-SimMUP).At last, vector operation part 62 offers recommendation information output 49 to each program of determining as programs recommended.In general, the predetermined number that all is confirmed as programs recommended program is set to 3.

In addition, if program vector PP is grouped, 62 bases of vector operation part so are about programs recommended information, one group of catalogue to recommending in user profile catalogue part 63 with first priority, and each program commending of one group recommending with first priority is with the first priority recommend programs group.

In addition, vector operation part 62 can be kept at theme in the initial directory stores part 45 by use, program vector PP is carried out filtration treatment, thereby produce the user model vector, and in user profile catalogue part 63, the user model vector that produces is made a catalogue, so that carry out matching treatment.User model will describe in detail in the back.

According to information as the initial catalogue of user in the initial directory stores part 45, information from initial directory stores part 45 receptions, from the positive history vectors UP of positive historical storage part 47 receptions or the negative history vectors MUP that receives from negative historical storage part 48, user profile catalogue part 63 produces user side effect vector EfUP and user side's negative effect vector EfMUP, preserves user side effect vector EfUP and user side's negative effect vector EfMUP.User side's effect vector EfUP of program vector PP is the vector that the component of program vector PP is provided with weighted value.The weighted components of program vector PP all is to will be to the important sport of the selection of user's recommend programs.In the selection of program, only use the weighting project of program vector PP.As a kind of alternative, user side's effect vector EfUP discloses the vector of user to the hobby of each project.On the other hand, user side's negative effect of program vector PP vector EfMUP is the vector that the component of program vector PP is provided with weighted value.But the weighted components of program vector PP all is to will be to the unessential sport of the selection of user's recommend programs.In the selection of program, only use the non-weighting project of program vector PP.As a kind of alternative, user side's negative effect vector EfMUP discloses the vector disliked of user to each project.

In other words, user side's effect vector EfUP of program vector PP is that each project of indication user side effect vector EfUP has great vector to the influence of the coupling in the matching treatment between program vector PP and the positive history vectors UP.On the other hand, user side's negative effect of program vector PP vector EfMUP is that each project of indication user side negative effect vector EfMUP has great vector to the influence of the mismatch in the matching treatment between program vector PP and the negative history vectors MUP.

User side's effect vector EfUP and user side's negative effect vector EfMUP can be provided with by the user, perhaps are configured to pre-determined a certain value.As a kind of alternative, also can produce user side effect vector EfUP and user side's negative effect vector EfMUP according to the information of the initial catalogue of the user in the initial directory stores part 45.

More particularly, user side's effect vector EfUP of hypothetical program vector PP is the weighted value of expression about the component of program vector PP, and the weighted components of program vector PP all is to will be to the important sport of the selection of user's recommend programs.In this case, if type is the component important to the user, so for program vector PP={ title Tm, type Gm, time period Hm, broadcasting station Sm, performing artist Pm, playwright, screenwriter/author/producer Am, content Km}, user side's effect vector EfUP is configured to representative value { 1,5,1,1,1,1,1}.On the other hand, if performing artist and type all are the components important to the user, so user side's effect vector EfUP be configured to

representative value

1,3,1,1,5,1,1}.

Suppose that user side's effect vector EfUP discloses the vector of user to the hobby of each project, and the type sport is the Gm={ drama, variety, physical culture, film, music, children's programs/education, generally literature and art/documentary film, news/report, other.In this case, if general literature and art/documentary film is the program category that the user prefers, so user side's effect vector EfUP be configured to representative value 0,0,0,0,0,0,5,0,0}.

In addition, can be according to positive history vectors UP or negative history vectors MUP, perhaps the program counting by the user was watched in the fixed time period can produce user side's effect vector EfUP and user side's negative effect vector EfMUP.In addition, also can produce user side effect vector EfUP and user side's negative effect vector EfMUP to every type.By with reference to the flow chart shown in the figure 22-27, the method that produces user side effect vector EfUP and user side's negative effect vector EfMUP is described.

In addition, if desired, user profile catalogue part 63 is kept at the information that produces in the processing that vector operation part 2 carries out.The information that produces in the processing that vector operation part 2 is carried out comprises one group and the user model vector of recommending with first priority.

Matching treatment part 43 described above can also be under the situation of the processing of not getting rid of program (this program is considered to the program that the user dislikes), programs recommended by from those programs of the program vector PP that all has the higher similarity that aligns history vectors UP, selecting, only according to for example positive historical determine programs recommended.The program that all has the program vector PP of the higher similarity that aligns history vectors UP all is that the user wants the program watched in a hurry.On the other hand, the program that is considered to the program that the user dislikes all is the program that the user is unwilling to watch.

The programs recommended information that matching treatment part 43 provides in 49 pairs of programs recommended tabulations 50 of recommendation information output is made a catalogue, and this information is offered TV display device 11 or recording/reproducing apparatus 12.Programs recommended tabulation 50 has the structure that can remove from program commending treatment facility 10, and is used to preserve the programs recommended information that is provided by recommendation information output 49.By programs recommended information being kept in the programs recommended tabulation 50, according to the historical information of preserving so far, even for example use different TV receiving equipments 4, different TV display devices 11 or different recording/reproducing apparatus 12 also can carry out the processing such as programs recommended processing and automatic recording processing.

In addition, if desired, matching treatment part 43 also is connected with driver 51.Disk 71, CD 72, magneto optical disk 73 or semiconductor memory 74 are installed on the driver 51.If desired, driver 51 and disk 71, CD 72, magneto optical disk 73 or semiconductor memory 74 swap datas.

Below with reference to the flow chart shown in Figure 14, illustrate that the positive history vectors of the positive history vectors of generation of being undertaken by program commending treatment facility 10 and negative history vectors and negative history vectors produce to handle 1.

At first at step S71, Operation Log obtains part 46 the initial catalogue data of reading from initial directory stores part 45 is offered negative historical storage part 48, and negative historical storage part 48 by with reference to initial catalogue data, produces negative history vectors MUP subsequently.

Subsequently, at next step S72, Operation Log obtains part 46 and determines whether the initial catalogue data that is kept in the initial directory stores part 45 is changed.If the definite result who produces in the processing that step S72 carries out indicates initial catalogue data to be changed, handling process is returned step S71 so, and Operation Log obtains the processing that part 46 is carried out step S71 and S72 once more.

On the other hand, if the definite result who produces in the processing that step S72 carries out indicates initial catalogue data not to be changed, handling process enters step S73 so, and Operation Log obtains part 46 and determines whether to have supplied with Operation Log from TV display device 11 or recording/reproducing apparatus 12.If the definite result's indication that produces in the processing that step S73 carries out is not also supplied with Operation Log from TV display device 11 or recording/reproducing apparatus 12, handling process is returned step S72 so, and Operation Log obtains the processing that part 46 is carried out step S72 and S73 once more.

On the other hand, if the definite result's indication that produces in the processing that step S73 carries out has been supplied with Operation Log from TV display device 11 or recording/reproducing apparatus 12, handling process enters step S74 so, and Operation Log obtains part 46 and determines whether supplied operation log is positive history.If Operation Log is a recording operation for example, the program vector PP of related program is positive historical program vector in this operation so.On the other hand, if Operation Log is to wipe also not reproduced record data, the program vector PP of related program is negative historical program vector in this operation so.

If it is positive history that the definite result who produces in the processing that step S74 carries out indicates supplied operation log, handling process enters step S75 so, Operation Log obtains part 46 and extracts the program vector PP that is confirmed as program related the just historical Operation Log from data acquiring portion 41, and this program vector PP is offered positive historical storage part 47.Positive historical storage part 47 saves as this program vector PP additional positive historical.

Subsequently, at next step S76, positive historical storage part 47 obtains the summation of the positive historical program vector PP of detail items or sport, thereby produces positive history vectors UP.After the processing that step S76 carries out finished, handling process was returned step S72, the processing of repeating step S72 and subsequent step.

On the other hand, if it is not positive history that the definite result who produces in the processing that step S74 carries out indicates supplied operation log, promptly, if it is negative historical that the definite result who produces in the processing that step S74 carries out indicates supplied operation log, handling process enters step S77 so, Operation Log obtains part 46 and extracts the program vector PP that is confirmed as bearing program related the historical Operation Log from data acquiring portion 41, and this program vector PP is offered negative historical storage part 48.Negative historical storage part 48 saves as this program vector PP additional negative historical.

Subsequently, at next step S78, negative historical storage part 48 obtains the summation of the negative historical program vector PP of detail items or sport, thereby produces negative history vectors MUP.After the processing that step S78 carries out finished, handling process was returned step S72, the processing of repeating step S72 and subsequent step.

In general, sport is by compound positive history vectors UP={ title Tup, and type Gup, performing artist Pup, playwright, screenwriter/author/producer Aup, content (keyword) Kup} represent, and stipulate detailed project in each sport.In this case, in each positive history vectors UP of expression sport, the numeral of expression vector sum is affixed on each detail items that constitutes sport.For example, as shown in Figure 15, the positive history vectors UP of expression type sport is type Gup={ (drama-25), (variety-34), (physical culture-42), (film-37), (music-73), (children's programs/education-120), (general literature and art/documentary film-3), (news/report-5), (other-23) }, wherein the numeral of the summation of the positive history of representative is affixed on each detail items that constitutes this sport.

The vectorial available word of the sport of representative such as the title sport is represented.For example, the positive history vectors UP of expression title sport is title Tup={ (title 1-12), (title 2-3) ..., wherein the numeral of the summation of the positive history of representative is affixed on each word that constitutes this sport.Be similar to positive history vectors UP, in negative history vectors MUP, represent the numeral of summation to be affixed on each project.

In the compound positive history vectors UP of each shown in Figure 15 (also having unshowned compound negative history vectors MUP among Figure 15 certainly), the sport that is used as vector components is a title, type, performing artist, playwright, screenwriter/author/producer and content (keyword).In this case, the number of sport is less than the number of the sport of the formation program vector PP of earlier in respect of figures 5 explanations.But needless to say, compound positive history vectors UP can have the sport identical with program vector PP with compound negative history vectors MUP.

In the processing with reference to the flowchart text shown in Figure 14,, produce negative history vectors MUP in front according to the initial catalogue data before the input operation daily record.But, when data are initially made a catalogue, be used to select user's the information of liking program also can be made a catalogue, make also can before the input operation daily record, produce positive history vectors UP.In addition, can be not according to initial catalogue data, but only produce positive history vectors UP or negative history vectors MUP according to Operation Log.

Like this, positive history vectors UP and negative history vectors MUP can independently be produced and be preserved, thereby, can carry out the matching treatment that user's hobby is discerned on pin-point accuracy ground.

Be noted that and find positive history and negative historical this moment with higher accuracy.In front in the processing with reference to the flowchart text shown in Figure 14, produce positive history vectors UP with form corresponding to the summation of the program vector PP of the positive history of all items.By the same token, produce negative history vectors MUP with form corresponding to the summation of the program vector PP of the negative history of all items.But, also can produce the positive history vectors UP of type corresponding to the form of the summation of the program vector PP of the positive history of type.By the same token, also can produce the negative history vectors MUP of type corresponding to the form of the summation of the program vector PP of the negative history of type.

Depend on the type of broadcast program, have that wherein to have only the user be not that a certain performing artist who misses potter appears in the program, make the common situation that in recommend programs, can not correctly reflect user's hobby.More particularly, we suppose that the user likes drama, and only like introducing and do not perform drama, but as the variety show of variety show performing artist's comedian A.For such user, we suppose that also the number of times of watching variety show for example is 2: 8 with watching the ratio of the number of times of drama.In this case, if accumulation performing artist's positive history so in some cases, is compared with comedian A under the situation of not considering type, the performing artist B that frequently appears in the drama has high score in positive history vectors UP, although performing artist B is not the user favorite actor.For such situation, it is recommended to introduce the recorded program frequently appear at the performing artist B in the drama, rather than it is recommended to introduce the variety show of comedian A.In order to address this problem, positive historical to every type of accumulation, and be every type according to the positive history of accumulation and produce positive history vectors UP.By the same token, negative historical to every type of accumulation, and be every type according to the negative history of accumulation and produce negative history vectors MUP.

For example suppose that the user is the admirer of commentator A, thereby the user likes watching " commentary of the baseball game of the B of team that commentator A explains orally ".In this case, the information as the physical culture of type project is easy to overlapping for historical.On the other hand, the information as " the commentator A " of performing artist's project is difficult to overlapping for historical.Thereby in some cases, the commentary of the baseball game of the B of team that is explained orally by another commentator is recommended, rather than introduces the variety show of commentator A.In order to address this is that, each performing artist to be accumulated positive history, and produce positive history vectors UP for each performing artist according to the positive history of accumulation.By the same token, each performing artist is accumulated negative history, and bear history vectors MUP for each performing artist produces according to the negative history of accumulation.

, can in recommend programs, reflect user's hobby with higher accuracy, and can not lose the real Cup of tea thing of user each element-specific accumulated history by as mentioned above.

Thereby, matching treatment part 43 can verify positive history vectors UP and the program vector PP that provides between coupling, and the coupling between the program vector PP of negative history vectors MUP and supply, thereby, can produce the programs recommended information of correct reflection user's hobby.

With reference to the flow chart shown in Figure 16, the following basis that interpretation carried out is to the history of every type of accumulation, and the positive history vectors and the negative history vectors that produce positive history vectors UP and negative history vectors MUP produce processing 2.

The processing of carrying out at step S81-S84 is identical with the processing of carrying out with reference to the step S71-S74 of the flow chart of Figure 14 explanation in the above respectively.That is, Operation Log obtains part 46 initial catalogue data is offered negative historical storage part 48, and negative historical storage part 48 by with reference to initial catalogue data, produces negative history vectors MUP subsequently.Afterwards, Operation Log obtains part 46 and determines whether initial catalogue datas are changed.

If initial catalogue data is not changed, Operation Log obtains part 46 and determines whether Operation Logs are provided so.

If Operation Log is provided, handling process enters step S84 so, determines whether Operation Log is positive history.If Operation Log is positive history, handling process enters step S85 so, Operation Log obtains part 46 and extracts and be confirmed as the program vector PP of program correspondence related the positive historical Operation Log from data acquiring portion 41, and a program vector PP that obtains is offered positive historical storage part 47.Positive historical storage part 47 extracts the type of this program vector PP subsequently.

Subsequently, at next step S86, positive historical storage part 47 saves as the additional positive historical of the type that extracts to the program vector PP that obtains from data acquiring portion 41.

Subsequently, at next step S87, positive historical storage part 47 obtains the summation about the positive historical program vector PP of all detail items of the type relevant with the program vector PP of extra preservation or all sports, so that produce the positive history vectors UP of the type.After the processing that step S87 carries out finished, handling process was returned step S82, carries out the processing of step S82 and subsequent step once more.

On the other hand, if it is negative historical that the definite result who produces in the processing that step S84 carries out indicates Operation Log, handling process enters step S88 so, Operation Log obtains part 46 and extracts and be confirmed as the program vector PP of program correspondence related the negative historical Operation Log from data acquiring portion 41, and a program vector PP that obtains is offered negative historical storage part 48.Negative historical storage part 48 extracts the type of this program vector PP subsequently.

Subsequently, at next step S89, negative historical storage part 48 saves as the additional negative historical of the type that extracts to the program vector PP that obtains from data acquiring portion 41.

Subsequently, at next step S90, negative historical storage part 48 obtains the summation about the negative historical program vector PP of all detail items of the type relevant with the program vector PP of extra preservation or all sports, so that produce the negative history vectors MUP of the type.After the processing that step S90 carries out finished, handling process was returned step S82, carries out the processing of step S82 and subsequent step once more.

By carrying out above-mentioned processing, be every type of positive history vectors UP of generation and negative history vectors MUP.Thereby, can reflect user's hobby at positive history vectors UP and negative history vectors MUP with higher accuracy, and can not lose the real Cup of tea thing of user.Thereby, can produce programs recommended information about correct reflection user's hobby.

With reference to the flow chart shown in Figure 17, below interpretation for program vector PP wherein, positive history vectors UP and negative history vectors MUP have the situation of the form that comprises all detail items components that are arranged in 1 array, the matching treatment 1 of the identification and matching program that is carried out.

At first at step S101, program vector extracts part 42 and obtains a plurality of programs from data acquiring portion 41, the program vector PP of program of broadcasting in the section at the fixed time for example, and program vector PP offered the normalized part 61 that in matching treatment part 43, adopts.The component of a word that constitutes title and content is all represented in 61 pairs of conducts of normalized part, is included in component in the program vector PP of each reception and the component that is included in from the positive history vectors UP that positive historical storage part 47 obtains and carries out normalization.Subsequently, normalized part 61 offers vector operation part 62 to the normalization result.

More particularly, we suppose the program vector PP typical example such as the title Tm={Toukaidou-1 of reception, Mitsuya-1, ghost-story-1}.In this case, normalized part 61 is normalized to title Tm={Toukaidou:0.33, Mitsuya:0.33, ghost-story:0.33} to program vector PP.By carrying out normalized like this,, can make the summation that is applied to as the weighted value of the word of the project that constitutes program title equal 1 for each program.

Subsequently, at next step S102, the vector operation part 62 that adopts in matching treatment part 43 is calculated similarity SimUP by utilizing the equation (1) that provides previously with the form of the cosine distance between each program vector PP and the positive history vectors UP.

Subsequently, at next step S103, vector operation part 62 is by comparing similarity SimUP mutually, from having the vector of the highest similarity SimUP that aligns history vectors UP, extract the short distance program vector PP of predetermined number, check similarity SimUP, similarity SimUP all has been calculated as the cosine distance between each program vector PP and the positive history vectors UP in the processing that step S102 carries out.For example, vector operation part 62 extracts 10 program vector PP.

Subsequently, at next step S104, vector operation part 62 is by utilizing the equation (2) that provides previously, and the form of the cosine distance between the negative history vectors MUP that obtains with each program vector PP of extracting in the processing of carrying out at step S103 with from negative historical storage part 48 is calculated similarity SimMUP.

Subsequently, at next step S105, the difference that vector operation part 62 is calculated between similarity SimUP and the corresponding similarity SimMUP, the program vector PP (or EPG data) of the most similar program of the predetermined number that extraction begins from the program with maximum difference is as recommendation information, and recommendation information is offered recommendation information output 49.As previously mentioned, similarity SimUP is the cosine distance between program vector PP and the positive history vectors UP.By the same token, similarity SimMUP is the cosine distance between program vector PP and the negative history vectors MUP.For example, vector operation part 62 extracts the program vector PP (perhaps EPG data) of the program with maximum difference.Recommendation information is made a catalogue in recommendation information tabulation 50 subsequently, and is exported to television display equipment 11 and recording/reproducing apparatus 12.At last, finish the execution of the processing of this flowcharting.

By carrying out above-mentioned processing, in program vector PP, positive history vectors UP and negative history vectors MUP all have under the situation of the form that comprises all detail items components that are arranged in 1 array, can be according to the similarity SimUP between program vector PP and the positive history vectors UP, and the similarity SimMUP between program vector PP and the negative history vectors MUP, definite conform to user's hobby programs recommended.

By the flow chart shown in reference Figure 18, following interpretation matching treatment 2, wherein not that each word as vector components is carried out normalization, but the cosine distance is calculated in each sport, and the summation of the cosine distance that obtains calculating, as being used to determine programs recommended similarity SimUP or similarity SimMUP.

At first at step S111, program vector extracts part 42 and extracts the program vector PP of a plurality of programs from data acquiring portion 41, and program vector PP is offered the vector operation part 62 of employing in matching treatment part 43.An example of program is the program of broadcasting in predetermined a period of time.Subsequently, for each sport that constitutes compound (compound) the positive history vectors UP quote previously, vector operation part 62 calculate each program vector PP and the positive history vectors UP that reads from positive historical storage part 47 between the cosine distance.

At step S112, according to the equation that provides previously (4), vector operation part 62 is calculated the summation of each cosine distance of calculating to each project in the processing that step S111 carries out, thereby obtains similarity SimUP subsequently.

Subsequently, at next step S113, by mutual relatively similarity SimUP, vector operation part 62 is checked the similarity SimUP that obtains in the processing that step S112 carries out.Each similarity SimUP represents the summation of the cosine distance between program vector PP and the positive history vectors UP.Vector operation part 62 from having the vector of highest similarity SimUP, extracts the short distance program vector PP of predetermined number subsequently.The predetermined number of the program vector PP that extracts is generally 10.

Subsequently, at next step S114,62 pairs of vector operation parts constitute each sport of the compound negative history vectors MUP that quotes previously, calculate each program vector PP of in the processing that step S113 carries out, extracting and the negative history vectors MUP that reads from negative historical storage part 48 between the cosine distance.

Subsequently, at next step S115, according to the equation that provides previously (4), vector operation part 62 is calculated the summation of each cosine distance of calculating to each project in the processing that step S114 carries out, thereby obtains similarity SimMUP.

Subsequently, at next step S116, the difference that vector operation part 62 is calculated between similarity SimUP and the corresponding similarity SimMUP, the program vector PP (or EPG data) of the most similar program of the predetermined number that extraction begins from the program with maximum difference is as recommendation information, and recommendation information is offered recommendation information output 49.As previously mentioned, similarity SimUP is the cosine distance between program vector PP and the positive history vectors UP.By the same token, similarity SimMUP is the cosine distance between program vector PP and the negative history vectors MUP.For example, vector operation part 62 extracts the program vector PP (perhaps EPG data) of 3 programs with maximum difference.Recommendation information is made a catalogue in recommendation information tabulation 50 subsequently, and is exported to television display equipment 11 and recording/reproducing apparatus 12.At last, finish the execution of the processing of this flowcharting.

By carrying out above-mentioned matching treatment, not that each word as vector components is carried out normalization, but the cosine distance is calculated in each sport, the summation of the cosine distance that obtains calculating is as being used to determine programs recommended similarity SimUP or similarity SimMUP.Thereby can be according to the similarity between program vector PP and the positive history vectors UP, and the similarity between program vector PP and the negative history vectors MUP, that determines to conform to user's hobby is programs recommended, and is not subjected to the influence by the historical overlapping deviation that causes between the detailed elements that belongs to different sports.

In front in the matching treatment 2 with reference to the matching treatment 1 of the flowchart text shown in Figure 17 and the flowchart text shown in the earlier in respect of figures 18, like best from user and to get rid of the program that the user by negative historical indication least likes the program by the highest similarity indication that aligns history vectors UP.But, positive historical by for example only utilizing, also can determine programs recommended.

In addition, in the processing of selecting program, in some cases, the user may like type to surpass the performing artist according to preferring news and report program, thinks perhaps that content is important and is indifferent to the performing artist, the project of having determined to be weighted He not being weighted.In other words, the user has some big events and the inessential project that is used to select program.

Thereby, the program side effect vector EfPP that matching treatment 1 and matching treatment 2 can use the front to illustrate, user side's effect vector EfUP and user side's negative vector EfMUP.In addition, can allow the user to determine whether to use program side effect vector EfPP, user side's effect vector EfUP and user side's negative vector EfMUP.

By the flow chart shown in reference Figure 19, following interpretation allows the user to determine whether to use for it, program vector PP, positive history vectors UP and negative history vectors MUP are the program side effect vector EfPP with vector of the form that comprises all detail items components that are arranged in 1 array, the matching treatment 3 of user side's effect vector EfUP.

At first, at step S121, vector operation part 62 obtains program side effect vector EfPP from initial directory stores part 45, and the utilization of user side's effect vector EfUP or user side's negative vector EfMUP is provided with data.The user by manipulation operations importation 44, is provided with data to utilization and has imported initial catalogue storage area 45.Whether utilization is provided with data and is meant and is shown in the matching treatment by utilizing program side effect vector EfPP, and user side's effect vector EfUP or user side's negative vector EfMUP use the information of weighted value.

Subsequently, at next step S122, if desired, vector operation part 62 is read user side's effect vector EfUP from user profile catalogue part 63, and calculate cosine distance between each program vector PP and the positive history vectors UP, thereby obtain similarity SimUP according to equation given below (5).

SimUP = \frac{{epd}_{1} \cdot {eud}_{1} \cdot p_{1} \cdot u_{1} + {epd}_{2} \cdot {eud}_{2} \cdot p_{2} \cdot u_{2} + . . .}{| PP | | UP |} - - - (5)

Attention in equation (5), program vector PP=(p ₁, p ₂...), positive history vectors UP=(u ₁, u ₂...), program side effect vector EfPP=(epd ₁, epd ₂...) and user side's effect vector EfUP=(eud ₁, eud ₂...) and be considered to represent respectively program vector PP, positive history vectors UP, program side effect vector EfPP and user side's effect vector EfUP.In addition, in equation (5), also hypothetical program side effect vector EfPP and user side's effect vector EfUP are used.But one of program side effect vector EfPP and user side's effect vector EfUP are any can not be used.In this case, value 1 has replaced the program side effect vector EfPP or the user side's effect vector EfUP that are not used.

In addition, user side's effect vector EfUP can be can be by the vector of user's setting, the vector that the initial setting up that provides according to the user is set, the perhaps vector that produces in user profile catalogue part 63.The back will describe the generation of user side's effect vector EfUP in detail with reference to the flow chart shown in the figure 22-25.

Subsequently, at next step S123, vector operation part 62 is checked the similarity SimUP that obtains by comparing similarity SimUP mutually in the processing that step S122 carries out.Each similarity SimUP represents the cosine distance between program vector PP and the positive history vectors UP.Vector operation part 62 from having the vector of highest similarity SimUP, extracts the short distance program vector PP of predetermined number subsequently.The predetermined number of the program vector PP that extracts is generally 10.

Subsequently, at next step S124, if desired, vector operation part 62 is read user side's negative effect vector EfMUP from user profile catalogue part 63, and each program vector PP that calculating is extracted according to equation given below (6) and the cosine distance between the negative history vectors MUP, thereby obtain similarity SimMUP in the processing that step S123 carries out.

SimMUP = \frac{{epd}_{1} \cdot {emd}_{1} \cdot p_{1} \cdot m_{1} + {epd}_{2} \cdot {emd}_{2} \cdot p_{2} \cdot m_{2} + . . .}{| PP | | MP |} - - - (6)

Attention in equation (6), program vector PP=(p ₁, p ₂...), negative history vectors MUP=(m ₁, m ₂...), program side effect vector EfPP=(epd ₁, epd ₂...) and user side's negative effect vector EfMUP=(emd ₁, emd ₂...) and be considered to represent respectively program vector PP, negative history vectors MUP, program side effect vector EfPP and user side's negative effect vector EfMUP.In addition, in equation (6), also hypothetical program side effect vector EfPP and user side's negative effect vector EfMUP is used.But one of program side effect vector EfPP and user side's negative effect vector EfMUP is any can not be used.In this case, value 1 has replaced the program side effect vector EfPP that is not used or user side's negative effect vector EfMUP.

In addition, user side's negative effect vector EfMUP can be can be by the vector of user's setting, the vector that the initial setting up that provides according to the user is set, the perhaps vector that produces in user profile catalogue part 63.The back will describe the generation of user side's negative effect vector EfMUP in detail with reference to the flow chart shown in Figure 26 or 27.

Subsequently, at next step S125, the difference that vector operation part 62 is calculated between similarity SimUP and the corresponding similarity SimMUP, the program vector PP (or EPG data) of the most similar program of the predetermined number that extraction begins from the program with maximum difference is as recommendation information, and recommendation information is offered recommendation information output 49.As previously mentioned, similarity SimUP is the cosine distance between program vector PP and the positive history vectors UP.By the same token, similarity SimMUP is the cosine distance between program vector PP and the negative history vectors MUP.For example, vector operation part 62 extracts the program vector PP (perhaps EPG data) of 3 programs with maximum difference.Recommendation information is made a catalogue in recommendation information tabulation 50 subsequently, and is exported to television display equipment 11 and recording/reproducing apparatus 12.At last, finish the execution of the processing of this flowcharting.

By carrying out above-mentioned matching treatment, according to data are set, program side effect vector EfPP, user side's effect vector EfUP or user side's negative effect vector EfMUP are used to extract recommendation information.Thereby, can recommend correctly to reflect the information of user's hobby.

In the above in the processing with reference to the flowchart text shown in Figure 19, program vector PP, positive history vectors UP and negative history vectors MUP are the vectors with the form that comprises all detail items components that are arranged in 1 array.But, can handle program vector PP to each sport, positive history vectors UP and negative history vectors MUP.

By the flow chart shown in reference Figure 20, following interpretation allows in the matching process for each sport, reflection program side effect vector EfPP, user side's effect vector EfUP or user side's negative effect vector EfMUP.

At first,, carry out the processing identical, thereby the utilization that obtains program side and user side's effect vector is provided with data with the step S121 of the flow chart shown in Figure 19 at step S131.

Subsequently, at next step S132, for each sport that constitutes compound positive history vectors UP, vector operation part 62 calculate the program vector PP of each input and the positive history vectors UP that reads from positive historical storage part 47 between the cosine distance.In this calculates, do not use effect vector.

Subsequently, at next step S133, according to equation given below (7), vector operation part 62 multiply by effect vector (words that need) to the cosine distance of calculating about each project, and calculate the cosine distance of each project or the summation of resulting product, thereby obtain similarity SimUP.

SimUP＝epd _t·eud _t·cosθu _t+epd _g·eud _g·cosθu _g

+epd _p·eud _p·cosθu _p+epd _a·eud _a·cosθu _a

+epd _k·eud _k·cosθu _k (7)

Note in the superincumbent equation (7) program vector PP=(p _t, p _g, p _p, p _a, p _k) and positive history vectors UP=(u _t, u _g, u _p, u _a, u _k) between cosine distance by (cos θ u _t, cos θ u _g, cos θ u _p, cos θ u _a, cos θ u _k) expression, program side effect vector EfPP is by EfPP=(epd _t, epd _g, epd _p, epd _a, epd _k) expression, and user side's effect vector EfUP is by EfUP=(eud _t, eud _g, eud _p, eud _a, eud _k) expression.In equation (7), program side effect vector EfPP and user side's effect vector EfUP are used in addition.But one of program side effect vector EfPP and user side's effect vector EfUP are any can not be used.In this case, value 1 replaces untapped program side effect vector EfPP or user side's effect vector EfUP.

Subsequently, at next step S134, vector operation part 62 is checked the form with the summation of the cosine distance between program vector PP and the positive history vectors UP, the similarity SimUP that obtains by comparing similarity SimUP mutually in the processing that step S133 carries out.Vector operation part 62 from having the vector of highest similarity SimUP, extracts the short distance program vector PP of predetermined number subsequently.The predetermined number of the program vector PP that extracts is generally 10.

Subsequently, at next step S135, vector operation part 62 is about each sport, calculate each program vector PP of in the processing that step S134 carries out, extracting and the negative history vectors MUP that reads from negative historical storage part 48 between the cosine distance.In this case, do not use effect vector.

Subsequently, at next step S136, according to equation given below (8), vector operation part 62 multiply by effect vector (words that need) to the cosine distance of calculating about each project, and the summation of calculating cosine distance or resulting product, thereby obtain similarity SimMUP.

SimMUP＝epd _t·emd _t·cosθm _t+epd _g·emd _g·cosθm _g

+epd _p·emd _p·cosθm _p+epd _a·emd _a·cosθm _a

+epd _k·emd _k·cosθm _k (8)

Note in the superincumbent equation (8) program vector PP=(p _t, p _g, p _p, p _a, p _k) and negative history vectors MUP=(m _t, m _g, m _p, m _a, m _k) between cosine distance by (cos θ m _t, cos θ m _g, cos θ m _p, cos θ m _a, cos θ m _k) expression, program side effect vector EfPP is by EfPP=(epd _t, epd _g, epd _p, epd _a, epd _k) expression, and user side's negative effect vector EfMUP is by EfMUP=(emd _t, emd _g, emd _p, emd _a, emd _k) expression.In equation (8), program side effect vector EfPP and user side's negative effect vector EfMUP are used in addition.But one of program side effect vector EfPP and user side's negative effect vector EfMUP is any can not be used.In this case, value 1 replaces untapped program side effect vector EfPP or user side's negative effect vector EfMUP.

Subsequently, at next step S137, the difference that vector operation part 62 is calculated between similarity SimUP and the corresponding similarity SimMUP, the program vector PP (or EPG data) of the most similar program of the predetermined number that extraction begins from the program with maximum difference is as recommendation information, and recommendation information is offered recommendation information output 49.As previously mentioned, similarity SimUP is the cosine distance between program vector PP and the positive history vectors UP.By the same token, similarity SimMUP is the cosine distance between program vector PP and the negative history vectors MUP.For example, vector operation part 62 extracts with maximum difference 3 the program vector PP (perhaps EPG data) of similar program.Recommendation information is made a catalogue in recommendation information tabulation 50 subsequently, and is exported to television display equipment 11 and recording/reproducing apparatus 12.At last, finish the execution of the processing of this flowcharting.

In above-mentioned processing, for each sport, effect vector all is used as weighted value.Thereby, can produce the recommendation information that conforms to the details of user's hobby.

By the flow chart shown in reference Figure 21, below interpretation carried out pass through to use to be produced by the positive history vectors of the flowchart text shown in the earlier in respect of figures 16 and negative history vectors and handle 2 positive history vectors UP and negative history vectors MUP about every type of generation, and use the matching treatment 5 of coming the identification and matching program for every type of user side's effect vector EfUP that provides and the vectorial EfMUP of user side's negative effect.

At first,, carry out the processing identical, thereby obtain the data that are provided with of effect vector with the step S121 of the flow chart shown in Figure 19 at step S141.

Subsequently, at next step S142, vector operation part 62 extracts the type of input program vector PP.In the following description, the type of input program vector PP for example is a drama.

Subsequently, at next step S143, vector operation part 62 is read positive history vectors UP based on the drama type from positive historical storage part 47, and about constituting each sport of compound positive history vectors UP, calculates the cosine distance between each input program vector PP and the positive history vectors UP.In this case, do not use effect vector.

Subsequently, at next step S144, according to equation given below (9), vector operation part 62 multiply by the cosine distance of calculating about each project user side's effect vector (words that need) of drama type, and calculate the cosine distance of each project or the summation of resulting product, thereby obtain similarity SimUP.

SimUP＝epd _t·eud _td·cosθu _td+epd _g·eud _gd·cosθu _gd

+epd _p·eud _pd·cosθu _pd+epd _a·eud _ad·cosθu _ad

+epd _k·eud _kd·cosθu _kd (9)

Note in the superincumbent equation (9) program vector PP=(p _t, p _g, p _p, p _a, p _k) and positive history vectors UP=(u _t, u _g, u _p, u _a, u _k) between cosine distance by (cos θ u _Td, cos θ u _Gd, cos θ u _Pd, cos θ u _Ad, cos θ u _Kd) expression, program side effect vector EfPP is by EfPP=(epd _t, epd _g, epd _p, epd _a, epd _k) expression, and user side's effect vector EfUP is by EfUP=(eud _Td, eud _Gd, eud _Pd, eud _Ad, eud _Kd) expression.In equation (9), program side effect vector EfPP and user side's effect vector EfUP are used in addition.But one of program side effect vector EfPP and user side's effect vector EfUP are any can not be used.In this case, value 1 replaces untapped program side effect vector EfPP or user side's effect vector EfUP.

Subsequently, at next step S145, vector operation part 62 is checked the form with the summation of the cosine distance between program vector PP and the positive history vectors UP, the similarity SimUP that obtains by comparing similarity SimUP mutually in the processing that step S144 carries out.Vector operation part 62 from having the vector of highest similarity SimUP, extracts the short distance program vector PP of predetermined number subsequently.The predetermined number of the program vector PP that extracts is generally 10.

Subsequently, at next step S146, vector operation part 62 is about each sport, calculate each program vector PP of in the processing that step S145 carries out, extracting and the negative history vectors MUP that reads from negative historical storage part 48 about the drama type between the cosine distance.In this case, do not use effect vector.

Subsequently, at next step S147, by using equation given below (10), vector operation part 62 multiply by the cosine distance of calculating about each project the effect vector (words that need) of drama type, and the summation of calculating cosine distance or resulting product, thereby obtain similarity SimMUP.

SimMUP＝epd _t·emd _td·cosθm _td+epd _g·emd _gd·cosθm _gd

+epd _p·emd _pd·cosθm _pd+epd _a·emd _ad·cosθm _ad

+epd _k·emd _kd·cosθm _kd (10)

Note in the superincumbent equation (10) program vector PP=(p _t, p _g, p _p, p _a, p _k) and negative history vectors MUP=(m _t, m _g, m _p, m _a, m _k) between cosine distance by (cos θ m _Td, cos θ m _Gd, cos θ m _Pd, cos θ m _Ad, cos θ m _Kd) expression, program side effect vector EfPP is by EfPP=(epd _t, epd _g, epd _p, epd _a, epd _k) expression, and user side's negative effect vector EfMUP is by EfMUP=(emd _Td, emd _Gd, emd _Pd, emd _Ad, emd _Kd) expression.In addition, in equation (10), program side effect vector EfPP and user side's negative effect vector EfMUP are used.But one of program side effect vector EfPP and user side's negative effect vector EfMUP is any can not be used.In this case, value 1 replaces untapped program side effect vector EfPP or user side's negative effect vector EfMUP.

Subsequently, at next step S148, the difference that vector operation part 62 is calculated between similarity SimUP and the corresponding similarity SimMUP, the program vector PP (or EPG data) of the most similar program of the predetermined number that extraction begins from the program with maximum difference is as recommendation information, and recommendation information is offered recommendation information output 49.As previously mentioned, similarity SimUP is the cosine distance between program vector PP and the positive history vectors UP.By the same token, similarity SimMUP is the cosine distance between program vector PP and the negative history vectors MUP.For example, vector operation part 62 extracts the program vector PP (perhaps EPG data) of 3 programs with maximum difference.Recommendation information is made a catalogue in recommendation information tabulation 50 subsequently, and is exported to television display equipment 11 and recording/reproducing apparatus 12.At last, finish the execution of the processing of this flowcharting.

In above-mentioned processing, for each sport, calculate the cosine distance between the positive history vectors UP of each program vector PP and type, and the cosine distance between the negative history vectors MUP of each program vector PP and type, by the effect vector of type is used as weighted value, obtain similarity.Thereby, can produce the recommendation information that conforms to the details of user's hobby.

In addition, according to the data of the initial catalogue of user in the initial directory stores part 45, produce user side effect vector EfUP and user side's negative effect vector EfMUP.As a kind of alternative, according to positive history vectors UP or negative history vectors MUP, perhaps the program counting by the user is watched in the given time produces the exclusive user side's effect vector EfUP of user and user side's negative effect vector EfMUP.

By the flow chart shown in reference Figure 22, the following program that passes through the user was watched in the set time that interpretation carried out is counted, and the user side's effect vector that produces user side's effect vector EfUP produces handles 1.

At first, at step S151, the user profile catalogue part 63 that adopts in matching treatment part 43 is selected untreated sport.

Subsequently, at next step S152, user profile catalogue part 63 is by with reference to the positive history that is kept in the positive historical storage part 47, detect the user at the fixed time, 1 week for example, the program of watching in January or March, request program vector extract part 42 and extract the program vector PP of the program that users watch in the given time from data acquiring portion 41, and about each detail items of being formed in the sport of selecting in the processing that step S151 carries out number counting to program.

More particularly, for example suppose that the sport of selecting is a type Gm={ drama in the processing that step S151 carries out, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other.In this case, be included in the project that constitutes this sport the number counting of the program that 63 couples of users of user profile catalogue part watch in the given time by the component of handle corresponding to the program vector PP of program.The number of for example supposing the program that the user watches in the given time is 50.In this case, the result of the number of the program that the user is watched in the given time counting is Gm={10 for example, 18,5,2,8,1,0,1, and 5}.

Subsequently, at next step S153, user profile catalogue part 63 request program vector extract parts 42 and extract the program vector PP of all programs of broadcasting from data acquiring portion 41 in the identical scheduled time, and about each detail items of being formed in the sport of selecting in the processing that step S151 carries out number counting to program.

More particularly, for example suppose that the sport of selecting is a type Gm={ drama in the processing that step S151 carries out, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other.In this case, be included in the project that constitutes this sport the number counting of 63 pairs of all programs of in the identical scheduled time, broadcasting of user profile catalogue part by the component of handle corresponding to the program vector PP of program.The number of for example supposing all programs of broadcasting in the identical scheduled time is 1000.In this case, be Gm={104 for example to the result of the number counting of all programs of in the identical scheduled time, broadcasting, 239,68,25,78,91,60,254,81}.

Subsequently, at next step S154, the ratio of the number of the actual program of watching of user profile catalogue part 63 calculating users and the number of all programs.As mentioned above, the number of the number of the actual program of watching of user and all programs obtains in the processing that step S152 and S153 carry out respectively.

The influence of audience ratings competition makes the program arrangement be counted as reflecting the arrangement of the hobby of general public.That is, in other words the processing of the ratio of the number of the actual program of watching of calculating user and the number of all programs is equal to by the number of all programs is used as master pattern, makes the normalized processing of number of the actual program of watching of user.The normalized vector that obtains in the processing that step S154 carries out is called as normalized vector D.

For example suppose that the sport of selecting is a type Gm={ drama in the processing that step S151 carries out, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other.The number of for example also supposing all programs of broadcasting in 1 week is (8,12,3,7,6,4,2,8,10), and the number of the actual program of watching of user is (4,0,1,2,3,4,5,5,2).In this case, normalized vector D=(4/8,0/12,1/3,2/7,3/6,4/4,1/2,2/8,2/10)=(0.5,0,0.33,0.28,0.5,1.0,0.5,0.13,0.2).Thereby the component that equals 1.0 normalized vector D means that all program users that broadcast in the given time about this component have watched.On the other hand, the component that equals 0 normalized vector D means that all program users that broadcast in the given time about this component do not watch.

Subsequently, at next step S155, user profile catalogue part 63 produces the effect vector of sport according to the result of calculation that obtains in the processing that step S154 carries out.

In order to produce effect vector, sport type Gm={ drama, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other a project be used to the value of setting up standard.For example suppose domestic consumer approximately watch one the week in broadcasting drama programs 20%.In this case, standard value is configured to 0.2.Because the effect vector of this sport is calculated as relative value, effect vector can have the value in the 0-1 scope.Thereby user side's effect vector has from normalized vector D that obtains as the result of the processing of carrying out at step S154 and the relative value that calculates as the value of setting of benchmark.

Thereby the effect vector E of expression user's the sport type Gm that likes type is calculated as E=(0.3 ,-0.2,0.13,0.08,0.3,0.8,0.3,-0.07,0.0), this obtains indicating the user to like children's programs/education type, but dislikes definite result of variety type.

Subsequently, at next step S156, user profile catalogue part 63 determines whether to produce the effect vector of all sports.If the definite result's indication that produces in the processing that step S156 carries out does not also produce the effect vector of all sports, handling process is returned step S151 so, once more the processing of execution in step S151 and subsequent step.

On the other hand, if the definite result's indication that produces in the processing that step S156 carries out has produced the effect vector of all sports, handling process enters step S157 so, and user profile catalogue part 63 is preserved the effect vector of all sports.At last, end is by the execution of the processing of this flowcharting.

By carrying out above-mentioned processing, can access the difference between the exclusive hobby of the hobby of general public and user.In addition,, for example March or half a year, recomputate user side's effect vector EfUP, can recommend to reflect the program of user's hobby in real time by about each predetermined a period of time.

In addition, in the processing with reference to the flowchart text shown in Figure 22, at preset time, for example 1 week, the program of watching in January or March obtains user side's effect vector EfUP according to the user in the above.As the program that calculates user side's effect vector EfUP can be in a plurality of periods, for example than short time interval, and medium period and than the program of watching in the long duration.In this case, the program of watching in each period according to the user obtains user side's effect vector EfUP, and user side's effect vector EfUP is used to determine recommendation information.

In above-mentioned processing, the exclusive hobby of user is used as user side's effect vector EfUP.But the exclusive hobby of user also can be used as the positive history vectors UP of matching treatment.

In addition, replace the program of all broadcasting, only the user is mainly watched the number counting of the program of broadcasting in the specific scheduled time slot of program.The example of this specific period is the so-called prime time from 18:00 to 22:00.By so specific scheduled time slot is set, can significantly reduce the quantity of the computing of the hobby of finding out general public.

With reference to the flow chart shown in Figure 23, below interpretation carried out pass through calculating cosine distance, obtain being used to using in matching treatment user side's effect vector of user side's effect vector EfUP of the difference between the hobby of exclusive hobby of user and general public to produce and handle 2, described cosine is apart from the similarity between the hobby of positive history vectors UP of expression and general public.

At first, at step S161, the positive history vectors UP that the user profile catalogue part 63 that adopts in matching treatment part 43 obtains to be kept in the positive historical storage part 47.

Subsequently, at next step S162, the standard that user profile catalogue part 63 obtains the hobby of expression general public is liked vectorial APP.

It can be the vector that receives from Distributor 5 that standard is liked vectorial APP, perhaps because effect envoy's purpose arrangement of audience ratings competition is counted as reflecting the arrangement of the hobby of general public, produce the 1 identical mode of handling according to top user side's effect vector with reference to the flowchart text shown in Figure 22, the number of all programs of broadcasting in scheduled time slot can be counted, and if desired, can carry out normalized, so that the standard that obtains is liked vectorial APP.

In Distributor 5, by utilizing the investigation result of general audience ratings, perhaps adopt other method, can the generation standard like vectorial APP.

Subsequently, at next step S163, user profile catalogue part 63 is to each sport, calculates positive history vectors UP and standard and likes cosine distance between the vectorial APP.The cosine distance is long more, and the similarity that so positive history vectors UP and standard are liked between the vectorial APP is high more.

Subsequently, at next step S164, user profile catalogue part 63 produces user side's effect vector EfUP of each sport by the inverse of the cosine distance that obtains calculating in the processing that step S163 carries out.At last, end is by the execution of the processing of this flow chart representative.The inverse of cosine distance is big more, and the similarity that so positive history vectors UP and standard are liked between the vectorial APP is low more.

By carrying out above-mentioned processing, can access user side's effect vector EfUP of the difference between the exclusive hobby of reflection hobby of general public and user.Handle by utilizing this user side's effect vector EfUP to carry out program commending, can determine to emphasize difference programs recommended between the exclusive hobby of the hobby of general public and user.

In the superincumbent explanation, program vector PP and positive history vectors UP are the vectors that provides for each sport.But notice that program vector PP and positive history vectors UP also can be the vectors with the form that comprises all detail items components that are arranged in 1 array.In addition in this case, needless to say also can carry out identical processing.

In addition, the similarity that positive history vectors UP and standard are liked between the vectorial APP not only is used to calculate effect vector, and is the designator of expression user's uniqueness, and such designator can be directly used in the recommendation of program.For example, like higher similarity between the vectorial APP, can at first recommend style to cater to the public's new program for positive history vectors UP and standard.

As above described with reference to the flow chart shown in Figure 22 and 23, if according in the learning process, the history of the operation that the user carries out obtains user side's effect vector EfUP rightly.But user side's effect vector EfUP also can initially be made a catalogue in advance.As another kind of alternative, the value that rule of thumb obtains also can be arranged to user side's effect vector EfUP in advance.

Be noted that by not only noting sport, and note constituting the detail items of each sport, can produce user side's effect vector EfUP.The performing artist Pm of for example supposition formation sport is divided into leading role performing artist and supporting role performing artist.Compare with the role assignments of leading role performing artist in drama or the film, the user who pays the utmost attention to supporting role performing artist's role assignments can be provided with user side's effect vector EfUP, to increase supporting role performing artist's weighted value.In another case, suppose the personnel that constitute as playwright, screenwriter/author/producer's element of sport, for example director, producer, author and photographer are distinguished mutually.In this case, compare with director and producer, the user who more appreciates photographer can be provided with user side's effect vector EfUP, so that increase photographer's weighted value.

In addition, can be every type and produce user side's effect vector EfUP, and be used in the matching treatment, contrast the positive history vectors UP and the program vector PP that produce the type of user side's effect vector EfUP into it, as the situation of top matching treatment 5 with reference to the flowchart text shown in Figure 21.

With reference to the flow chart shown in Figure 24, following passing through that interpretation carried out about every type, to the program counting that the user watches in scheduled time slot, the user side's effect vector that produces user side's effect vector EfUP produces handles 3.

At first, at step S171, the user profile catalogue part 63 that adopts in matching treatment part 43 is selected will be about the type of its program counting that user is watched in scheduled time slot.

Subsequently, at next step S172, user profile catalogue part 63 is selected the sport of being untreated of selected type.

Subsequently, at next step S173, by with reference to being kept at positive history in the positive historical storage part 47, user profile catalogue part 63 at scheduled time slot, 1 week for example, detects the program of selected type the user in those programs of watching in January or March.Subsequently, user profile catalogue part 63 request program vector extract part 42 extracts detected program from data acquiring portion 41 program vector PP, and about being formed in each detail items of the type sport of selecting in the processing that step S172 carries out, to detected program counting.

Subsequently, at next step S174, user profile catalogue part 63 request program vector extract part 42 and extract as the program that belongs to selected type from data acquiring portion 41, the program vector PP of all programs that the user watches in scheduled time slot, and about being formed in each detail items of the type sport of selecting in the processing that step S172 carries out, to such program counting.

Subsequently, at next step S175, user profile catalogue part 63 by the count results that in the processing that step S173 carries out, obtains divided by the count results that obtains in the processing of carrying out at step S174, calculate the ratio of number of all programs of the number of the actual program of watching of user and selected type.

As mentioned above, envoy's purpose arrangement that influences of audience ratings competition is counted as the arrangement that reflects that the public likes.That is, in other words the processing of the ratio of the number of the number of the actual program of watching of calculating user and all programs of selected type is equal to by the number of all programs is used as master pattern, makes the normalized processing of number of the actual program of watching of user.The normalized vector that obtains in the processing that step S175 carries out is called as normalized vector D '.

For example suppose drama from sport type Gm={, variety, physical culture, film, music, children's programs/education, general literature and art/documentary film, news/report, other middle drama type of selecting.For example also suppose sport time period Tm={ morning for the program vector PP corresponding with the drama type of selecting, daytime, at dusk, prime time, the late into the night }, the number of all programs of broadcasting is (10,35,7,53,17) in 1 week, the number of the actual program of watching of user is (5,0,0,8,4).In this case, normalized vector is D '=(5/10,0/35,0/7,8/53,4/17)=(0.5,0,0,0.28,0.15,0.24).Thereby the component that equals 1.0 normalized vector D ' means that all program users that broadcast in the given time about this component have watched.On the other hand, the component that equals 0 normalized vector D ' means that all program users that broadcast in the given time about this component do not watch.

Subsequently, at next step S176, user profile catalogue part 63 produces the effect vector of the sport of selected type according to the result of calculation that obtains in the processing that step S175 carries out.

In order to produce effect vector, sport time period Tm={ morning, daytime, at dusk, prime time, the late into the night } a project be used to the value of setting up standard.For example suppose domestic consumer approximately watch inherent prime time broadcasting of a week drama programs 20%.In this case, standard value is configured to 0.2.Because the effect vector of this sport is calculated as relative value, effect vector can have the value in the 0-1 scope.Thereby user side's effect vector has from normalized vector D ' that obtains as the result of the processing of carrying out at step S175 and the relative value that calculates as the value of setting of benchmark.

Thereby the effect vector E ' of expression user's the sport type Gm that likes type is calculated as E '=(0.3 ,-0.2,-0.2 ,-0.05,0.04), this obtains indicating the user to like the drama of period between morning, but dislikes definite result of the drama of daytime and the dusk period.

Subsequently, at next step S177, user profile catalogue part 63 determines whether to produce the effect vector of all sports about selected type.If the definite result's indication that produces in the processing that step S177 carries out does not also produce the effect vector of all sports about selected type, handling process is returned step S172 so, once more the processing of execution in step S172 and subsequent step.

On the other hand, if the definite result's indication that produces in the processing that step S177 carries out has produced the effect vector of all sports about selected type, handling process enters step S178 so, and user profile catalogue part 63 determines whether to have handled all types.If all types is not also handled in the definite result's indication that produces in the processing that step S178 carries out, handling process is returned step S171 so, once more the processing of execution in step S171 and subsequent step.

On the other hand, if all types has been handled in the definite result's indication that produces in the processing that step S178 carries out, handling process enters step S179 so, and user profile catalogue part 63 is preserved the effect vector of all sports, and the execution of end process.

By carrying out above-mentioned processing, can access about every type the difference between the exclusive hobby of the hobby of general public and user.In addition, by about each predetermined a period of time, for example March or half a year, recomputate user side's effect vector EfUP, can recommend to reflect the program of user's hobby in real time, as the situation of the processing of the flowchart text shown in the earlier in respect of figures 22.

In addition, in the processing with reference to the flowchart text shown in Figure 24, at preset time, for example 1 week, the program of watching in January or March obtains user side's effect vector EfUP according to the user in the above.But, can be in a plurality of periods as the program that calculates user side's effect vector EfUP, for example than short time interval, medium period and than the program of watching in the long duration is as the situation of the processing of the flowchart text shown in the earlier in respect of figures 22.In this case, the program of watching in each period according to the user obtains user side's effect vector EfUP, and user side's effect vector EfUP is used to determine recommendation information.

In addition, in the processing with reference to the flowchart text shown in Figure 24, replace the program of all broadcasting in the above, only the user is mainly watched the program counting of broadcasting in the specific scheduled time slot of program.The example of this specific period is the so-called prime time from 18:00 to 22:00.

With reference to the flow chart shown in Figure 25, below interpretation carried out pass through calculating cosine distance, obtain to be used in matching treatment, using user side's effect vector of user side's effect vector EfUP of the difference between the hobby of exclusive hobby of user and general public to produce and handle 4, described cosine is apart from representing about every type the similarity between the hobby of positive history vectors UP and general public.

At first, at step S191, the user profile catalogue part 63 that adopts in matching treatment part 43 is selected to carry out the type of this processing to it.

Subsequently, at next step S192, the positive history vectors UP of user profile catalogue part 63 from be kept at positive historical storage part 47 obtains the positive history vectors UP of selected type.

Subsequently, at next step S193, the standard that user profile catalogue part 63 is liked vectorial APP acquisition selected type from the standard of the hobby of expression general public is liked vectorial APP.

As previously mentioned, it can be the vector that receives from Distributor 5 that standard is liked vectorial APP, perhaps because the effect of audience ratings competition makes the program arrangement be counted as reflecting the arrangement of the hobby of general public, produce the 3 identical modes of handling according to top user side's effect vector with reference to the flowchart text shown in Figure 24, the number of all programs of broadcasting in scheduled time slot of every type can be counted, and if desired, can carry out normalized, so that the standard that obtains is liked vectorial APP.

In Distributor 5,, perhaps, can the generation standard like vectorial APP by adopting other method by utilizing the investigation result of general audience ratings.

Subsequently, at next step S194, user profile catalogue part 63 is liked vectorial APP by positive history vectors UP and the standard of using each sport, and about each sport, the positive history vectors UP of calculating selected type and the standard of selected type are liked the cosine distance between the vectorial APP.The cosine distance is long more, and the similarity that so positive history vectors UP and standard are liked between the vectorial APP is high more.

Subsequently, at next step S195, user profile catalogue part 63 produces user side's effect vector EfUP of each sport by the inverse of the cosine distance that obtains calculating in the processing that step S194 carries out.The inverse of cosine distance is big more, and the similarity that so positive history vectors UP and standard are liked between the vectorial APP is low more.

Subsequently, at next step S196, user profile catalogue part 63 determines whether all types is processed.If all types is not also handled in the definite result's indication that produces in the processing that step S196 carries out, handling process is returned step S191 so, once more the processing of execution in step S191 and subsequent step.On the other hand, if the definite result who produces indicates all types processed, finish the execution of the processing of this flow chart representative so in the processing that step S196 carries out.

By carrying out above-mentioned processing, can access user side's effect vector EfUP of the difference between the exclusive hobby of every type the hobby of reflection general public and user.

In addition, according to the identical mode of processing of the flowchart text shown in the earlier in respect of figures 23, like the inverse of the calculating similarity between the vectorial APP by obtaining negative history vectors MUP and standard, user profile catalogue part 63 can produce user side's negative effect vector EfMUP of each sport.

With reference to the flow chart shown in Figure 26, the following hobby of passing through relatively more negative history vectors MUP and general public that interpretation carried out, the user side's negative effect vector that obtains user side's negative effect vector EfMUP produces handles 1.

At first, at step S201, the user profile catalogue part 63 that adopts in matching treatment part 43 obtains to be kept at the negative history vectors MUP in the negative historical storage part 48.

At next step S202, the standard that user profile catalogue part 63 obtains the hobby of expression general public is liked vectorial APP.

It can be the vector that receives from Distributor 5 that standard is liked vectorial APP, perhaps because the influence of audience ratings competition makes the program arrangement be counted as reflecting the arrangement of the hobby of general public, produce the 1 identical mode of handling according to top user side's effect vector with reference to the flowchart text shown in Figure 22, the number of all programs of broadcasting in scheduled time slot can be counted, and can carry out normalized if desired, so that the standard that obtains is liked vectorial APP.

At next step S203, user profile catalogue part 63 is to each sport, and negative history vectors MUP of calculating and standard are liked the cosine distance between the vectorial APP.The cosine distance is long more, and the similarity that so negative history vectors MUP and standard are liked between the vectorial APP is high more.

At next step S204, user profile catalogue part 63 produces user side's negative effect vector EfMUP of each sport by the inverse of the cosine distance that obtains calculating in the processing that step S203 carries out.Subsequently, finish the execution of the processing of this flow chart representative.

By carrying out above-mentioned processing, can produce user side's negative effect vector EfMUP.Thereby, can from programs recommended tabulation, get rid of the program that the user dislikes effectively.

With reference to the flow chart shown in Figure 27, the following user side's negative effect vector that obtains user side's negative effect vector EfMUP of every type that interpretation carried out produces handles 2.

At first, at step S211, the user profile catalogue part 63 that adopts in matching treatment part 43 is selected will be at subsequent step to its type of handling.

At next step S212, obtain the negative history vectors MUP of selected type among the negative history vectors MUP of user profile catalogue part 63 from be kept at negative historical storage part 48.

At next step S213, user profile catalogue part 63 obtains all to represent that the standard of the hobby of general public likes one of vectorial APP, likes vectorial APP as the standard of selected type.

Subsequently, at next step S214, user profile catalogue part 63 is liked vectorial APP according to negative history vectors MUP and standard, and the standard of calculating the negative history vectors MUP of selected type and selected type about each sport is liked the cosine distance between the vectorial APP.The cosine distance is long more, and the similarity that so negative history vectors MUP and standard are liked between the vectorial APP is high more.

Subsequently, at next step S215, user profile catalogue part 63 produces user side's negative effect vector EfMUP of each sport by the inverse of the cosine distance that obtains calculating in the processing that step S214 carries out.The inverse of cosine distance is big more, and the similarity that so negative history vectors MUP and standard are liked between the vectorial APP is low more.

Subsequently, at next step S216, user profile catalogue part 63 determines whether all types is processed.If all types is not also handled in the definite result's indication that produces in the processing that step S216 carries out, handling process is returned step S211 so, once more the processing of execution in step S211 and subsequent step.On the other hand, if the definite result who produces indicates all types processed, finish the execution of the processing of this flow chart representative so in the processing that step S216 carries out.

By carrying out above-mentioned processing, can produce user side's negative effect vector EfMUP about every type.Thereby, can from programs recommended tabulation, get rid of the program that the user dislikes effectively.

Be noted that also in the processing with reference to the flowchart text shown in Figure 23 and the 25-27 in front that the n form doubly of the cosine inverse distance that obtains about each sport obtains user side's effect vector EfUP and user side's negative effect vector EfMUP.As a kind of alternative, cast out the value that the numerical digit of predetermined number obtains according to the technology of rounding up from the inverse of cosine distance and also can be used as user side's effect vector EfUP or user side's negative effect vector EfMUP.As another kind of alternative,, can obtain user side's effect vector EfUP and user side's negative effect vector EfMUP as from the corresponding inverse of cosine distance, deducting 1 result.As another kind of alternative, can from the corresponding inverse of cosine distance, deduct 1 result's n form doubly, obtain user side's effect vector EfUP and user side's negative effect vector EfMUP.

In the superincumbent explanation, program vector PP and negative history vectors MUP are the vectors of each sport.But be noted that program vector PP and negative history vectors MUP also can be the vectors with the form that comprises all detail items components that are arranged in 1 array.In addition in this case, needless to say also can carry out identical processing.

In addition, program vector extracts the program vector PP that part 42 extracts can have additional group ID or cluster code thereon, and described group of ID or cluster code are represented corresponding to the group under the program of this program vector PP.As mentioned above, by carrying out the processing by the flowcharting shown in Fig. 7 and 8, generation group ID and cluster code respectively.

If the program that the user likes watching forms serial, and all programs that form this serial can reduce the treating capacity of recommending these programs so with recommended.By the program that the program commending of the phase of first priority on the same group enjoyed a lot for the user, also can reduce the quantity of recommendation process.

Below with reference to the flow chart shown in Figure 28, the matching treatment that comprises that group is recommended is described.

At first, at step S221, the vector operation part 62 that adopts in matching treatment part 43 determines whether the program vector PP that program vector extraction part 42 extracts has the information of a certain group of additional identification, so that determine whether the program relevant with this program vector PP belongs to this group.As previously mentioned, the example that appends to the information of a certain group of identification on the program vector PP is group ID and cluster code.

If the definite result who produces in the processing that step S221 carries out indicates program vector PP to have the information of a certain group of additional identification, handling process enters step S222 so, and vector operation part 62 determines whether the group of group ID or cluster code indication is categorized as the group of recommending with first priority in user profile catalogue part 63.

If the group that the definite result's indication that produces in the processing that step S222 carries out is represented by group ID or cluster code is categorized as the group of recommending with first priority in user profile catalogue part 63, handling process enters step S223 so, and the information of 62 programs relevant with program vector PP of vector operation part offers recommendation information output 49 as recommendation information.Recommendation information output 49 to the programs recommended catalogue of recommendation information suggestion, and is exported to television display equipment 11 or recording/reproducing apparatus 12 to programs recommended information in programs recommended tabulation 50.Subsequently, handling process enters the step S227 that illustrates later.

On the other hand, if the definite result who produces in the processing that step S221 carries out indicates program vector PP not have the information of a certain group of additional identification, the group that the definite result's indication that perhaps produces in the processing that step S222 carries out is represented by group ID or cluster code is not categorized as the group of recommending with first priority in user profile catalogue part 63, handling process enters step S224 so, carries out the front respectively with reference to one of any in the matching treatment 5 of the matching treatment 1 of the flowchart text shown in the figure 17-21.

Subsequently, at next step S225, by determining whether programs recommended program vector has the information of a certain group of additional identification, vector operation part 62 determines whether recommend programs belongs to this group in matching treatment 1 that step S224 carries out is one of any to matching treatment 5.As previously mentioned, the example that appends to the information of a certain group of identification on the program vector PP is group ID and cluster code.If the definite result who produces indicates recommend programs not belong to a certain group, finish the execution of the processing of this flow chart representative so in the processing that step S225 carries out.

On the other hand, if the definite result who produces in the processing that step S225 carries out indicates recommend programs to belong to a certain group, handling process enters step S226 so, and vector operation part 62 is kept at the group ID (perhaps cluster code) that adds program vector PP as the group of recommending with first priority in the user profile catalogue part 63.

After the processing that step S223 or S226 carry out finishes, handling process is carried out step S227, vector operation part 62 is kept at the daily record of the operation in the negative historical storage part 48 with reference to bearing historical operation as regarding as, determines whether to receive that being counted as negative historical operation imports.The operation checked in the processing that step S227 carries out input is about recommend programs in the processing of carrying out at step S223, perhaps as one group that recommends with first priority, and the program of in the processing that step S226 carries out, being made a catalogue and the operation input imported.This is because the program of being made a catalogue in the processing that step S226 carries out is that in the recommend programs has sorted out program in the processing that step S224 carries out.The example that is counted as the operation input of negative history is that the request that the user imports is watched another program or write down another program, because the user dislikes institute's recommend programs, perhaps before the reproduction of recorded program, asks the order of wiping this recorded program.

Be not counted as negative historical operation input if the definite result's indication that produces is received in the processing that step S227 carries out, finish the execution of the processing of this flow chart representative so.

On the other hand, if having received, the definite result's indication that produces is counted as negative historical operation input in the processing that step S227 carries out, handling process is carried out step S228 so, vector operation part 62 is kept in the catalogue in the user profile catalogue part 63 and removes the group that comprises the target program from as the catalogue with the first priority recommend programs group.

Be noted that by carrying out the title packet transaction 1 of the flowchart text shown in the earlier in respect of figures 9, in some cases, a plurality of groups of ID can be assigned to a program vector PP.Specifically, under these circumstances, in the processing that step S228 carries out, according to the input of negative historical operation, all group ID can be from the catalogue as the group of recommending with first priority, is kept in the catalogue in the user profile catalogue part 63 to be removed.As a kind of alternative, all being counted as negative historical operation input is accumulated, when the user imported the negative historical input operation of predetermined number, all group ID can be from the catalogue as the group of recommending with first priority, was kept in the catalogue in the user profile catalogue part 63 to be removed.In addition, for program vector PP wherein and the related correspondingly situation of group ID, can adopt this feature.

By carrying out above-mentioned processing, can recommend the user to like all programs of the serial watched, the program program on the same group that enjoys a lot with the user can be recommended to belong to, thereby the operational processes of recommendation can be reduced.

In comprising aforesaid group of matching treatment of recommending, also can be according to the user's operation history that obtains from television display equipment 11 or recording/reproducing apparatus 12, produce positive history vectors UP and negative history vectors MUP about the group of correspondence.

In addition, according to detected positive history the user's operation history that obtains from television display equipment 11 or recording/reproducing apparatus 12, the number of operation of reservation of watching identical serial or carrying out the record of serial can be counted, when resulting counting surpasses predetermined value, can first priority recommend this serial, and not carry out matching treatment specially.

In addition, replace carrying out historical programs recommended processing, can user model be set according to the theme that the user is provided with in advance, and, carry out programs recommended processing according to user model according to the user.

By via the filter process that utilizes theme, extract and the program of the theme correspondence that the user is provided with in advance the acquisition user model.Initial directory stores part 45 is preserved the theme of user's input.A plurality of themes can be stored in the initial directory stores part 45.In addition, the user can upgrade theme rightly by operating.Vector operation part 62 is from the program vector PP as the program that all is used to produce user model, and program vector extracts to extract in the program vector PP that part 42 supplies with and includes the program vector PP that is kept at the theme in the initial directory stores part 45.Subsequently, the summation of the program vector PP that vector operation part 62 obtains extracting, if desired, order makes described summation normalization, thereby produces the user model vector.At last, vector operation part 62 is kept at the user model vector in the user profile catalogue part 63.

For example we suppose that user's stipulated time section Hm=" after the 11PM " and type Gm=" variety " are as theme.In this case, produce user model " variety show in the late into the night ", carry out filtration treatment by using theme.User model " variety show in the late into the night " comprises the comedian who performs as the configuration key element in the variety show that 23:00 broadcasted in the time period in the late into the night afterwards.Thereby for example, in matching treatment, the program that this performing artist occurs is extracted from the program of the type different with the variety type, as programs recommended.The example of the type different with the variety type is the drama type.Thereby, and each project is carried out matching treatment comparing, the program of the limit of consideration of the type can be confirmed as user's hobby head and shoulders above.

The program of noting being used for producing user model generally can be all programs of broadcasting at the fixed time, perhaps at scheduled time slot, and one group of program of broadcasting in for example so-called prime time.

In addition, for example, under identical filtercondition, the different programs of respectively organizing is carried out filtration treatment, so that produce in detail corresponding to each different program arrangement, the user model vector of the program arrangement of the different time in 1 year and the program of different periods arrangement for example.

More particularly, even at the same terms, for example under " type Gm=music ", modern program arrangement has the airtime and the performing artist of the program arrangement that was different from before 10 years.Thereby, can produce " modern music fan " user model different with " music fans before 10 years " user model.Thereby, can recommend to like listening drama or the film of the pop singer of its song as the performer by its " music fans before 10 years ".In addition, can also recommend wherein modern singer to sing the program that " music fans before 10 years " like the first song listened to.

In addition, even for identical condition, for example " type Gm=music " also can carry out filtration treatment, thereby to obtain mainly be that the adult watches the period of program, and mainly is the many groups of different programs of children's period of watching program.Thereby, can produce different user models.In general, adult's period is from 20:00 to 24:00, and children's period is from 15:00 to 20:00.

Subsequently, the similarity between program vector PP that vector operation part 62 calculating program vector extraction parts 42 provide and the user model vector of in user profile catalogue part 63, making a catalogue.Subsequently, according to result of calculation, vector operations part 62 produces recommendation information, and this information is offered recommendation information output 49.Recommendation information output 49 is made a catalogue to recommendation information in programs recommended tabulation 50, and this information is offered television display equipment 11 or recording/reproducing apparatus 12.

Below with reference to the flow chart shown in Figure 29, the matching treatment of using user model is described.

At first, at step S231, the vector operation part 62 that adopts in matching treatment part 43 extracts the program vector PP that part 42 obtains to be used to produce the program of user model from the program vector that extracts program vector PP.The program that is used to produce user model can be in predetermined a period of time, for example the program of broadcasting in January or March.As a kind of alternative, the program that is used for producing user model can be the scheduled time slot at predetermined certain hour, for example the program of broadcasting in the prime time.As another kind of alternative, the program that is used to produce user model can be in the past, for example the program of broadcasting in the predetermined a period of time before 10 years or 20 years.

Subsequently, at next step S232, initial directory stores part 45 obtains the theme of user by 44 inputs of manipulation operations importation, and preserves this theme.Vector operation part 62 is read described theme from initial directory stores part 45, and filtercondition be arranged in this theme.For example, filtercondition can be described as hash table.

Subsequently, at next step S233, vector operation part 62 is carried out filtration treatment according to the filtercondition that is provided with to the program vector PP that obtains in the processing of carrying out at step S231 in the processing that step S232 carries out, so that extract the program vector PP that conforms to filtercondition.For example, the program that all is used to produce user model is all programs of broadcasting in the past 3 months.In this case, vector operation part 62 is carried out filtration treatment according to filtercondition " title Tm or type Gm=football " to the program vector PP in past 3 months.

Subsequently, at next step S234, the summation of the program vector PP of the filtration treatment output that vector operation part 62 obtains carrying out at step S233, and this summation as the user model vector.For example, vector operation part 62 is carried out filtration treatment according to filtercondition " title Tm or type Gm=football " to program vector PP.In this case, produce the user model vector of user model " football fan ".

Subsequently, at next step S235, the cosine distance between the program vector PP that the user model vector sum program vector extraction part 42 that 62 calculating of vector operation part produce in the processing that step S234 carries out extracts.Program vector PP is all as the program of candidate's recommendation information, that is, and and the program vector PP of the program that will broadcast after a while.

Subsequently, at next step S236, vector operation part 62 is by similarity relatively mutually, checks with as the program vector PP of the program of candidate's recommendation information and the form of the cosine distance between the user model vector, the similarity that obtains in the processing that step S235 carries out.According to check result, vector operation part 62 extracts the short distance program vector PP of predetermined number subsequently from having the vector of highest similarity, as recommendation information, and this recommendation information is offered recommendation information output 49.Recommendation information output 49 is made a catalogue to this recommendation information in programs recommended tabulation 50, and this information is offered television display equipment 11 or recording/reproducing apparatus 12, and end process.

By carrying out above-mentioned processing,, also can recommend the program that conforms to the theme of user's setting even the history of the operation that the user carries out does not exist.In addition, in the past, in the filtration treatment of carrying out according to filtercondition " title Tm or type Gm=football ", variety show or the drama acted the leading role by the football player are not extracted.On the other hand, in front in the processing with reference to the flowchart text shown in Figure 29, by user model being set according to theme " title Tm or type Gm=football ", variety show of being acted the leading role by the football player or drama can be extracted and be recommendation information, even this program and ineligible " title Tm or type Gm=football ".This is to comprise a large amount of football players because of content and performing artist as the project that constitutes the user model vector.

In above-mentioned processing, obtain cosine distance between the user model vector sum program vector PP as similarity.But also the form of the summation of each the cosine distance that can calculate about sport obtains similarity.

The processing that is noted that generation user model vector also can be carried out in Distributor 5.In this case, produce the program vector PP that part 23 produces, carry out the processing of the step S231-S234 of the flow chart shown in Figure 29 by using the program vector that generally illustrates by earlier in respect of figures 2.

In addition, except the recommendation of the program that conforms to user's hobby, by selecting to have the low similarity between program vector PP and the positive history vectors UP, and the program of the low similarity between program vector PP and the negative history vectors MUP, the user had both disliked the recommended probability of also inoffensive program and had become big.That is, the recommended probability of program with characteristic that the user do not watch in the past becomes big.In other words, in addition the recommended probability of program not attempting being refused by the user under the situation of program become big.In order to advance learning process by the hobby that extracts the user, importantly assessment has the program of the characteristic that the user do not watch in the past.

Thereby the recommendation results that presents to the user is endowed unpredictability.So, not only may cause the abundant of user interest, but also may obtain very important historical information concerning the program of recommending match user hobby better.

Below, with reference to the flow chart shown in Figure 30, illustrate that the programs recommended unusual programs recommended selection of the identification of being carried out unusual (exceptional) is handled.

At first, at step S241, the vector operation part 62 that adopts in matching treatment part 43 uses program vector to extract the program vector PP that part 42 provides, the negative history vectors MUP that is kept at the positive history vectors UP in the positive historical storage part 47 and is kept in the negative historical storage part 48 comes about each sport, calculate the cosine distance between program vector PP and the positive history vectors UP, and the cosine distance between program vector PP and the negative history vectors MUP.

Subsequently, at next step S242, vector operation part 62 obtains the summation of the cosine distance calculated about each project in positive history side and is bearing the summation of history side about the cosine distance of each project calculating.That is, the processing of carrying out at step S241 and S242 produces the similarity SimUP between program vector PP and the positive history vectors UP about each sport, and the similarity SimMUP between program vector PP and the negative history vectors MUP.

Subsequently, at next step S243, vector operation part 62 calculate the similarity represent between program vector PP and the positive history vectors UP lowly, and the low unusual recommendation of the similarity between program vector PP and the negative history vectors MUP.

Specifically, unusual recommendation can be by expression formula (1-SimUP) * (1-SimMUP) or (1/SimUP) * (1/SimMUP) statement.

At next step S244, vector operation part 62 is according to the result of calculation that is produced by the processing of carrying out at step S243 subsequently, and extraction all has the program of big unusual recommendation as recommendation information.At last, finish the execution of the processing of this flow chart representative.

By carrying out above-mentioned processing, the program with characteristic that the user do not watch so far can be extracted to programs recommended.Thereby, will be endowed unpredictability to the selection of user's recommend programs.Thereby, not only may cause the abundant of user interest, but also may obtain very important historical information concerning the program of recommending match user hobby better.

By every processing of carrying out illustrating with reference to figure 13-30 so far, program commending treatment facility 10 can produce the recommendation information that will be provided for television display equipment 11 or recording/reproducing apparatus 12.

Television display equipment 11 or recording/reproducing apparatus 12 also receive as satellite ripple or earthwave and are captured, and by the broadcast singal of TV receiving apparatus 4 decoding.

According to the operation that the user carries out, television display equipment 11 shows the broadcast singal that receives from TV receiving apparatus 4, perhaps the reproduction data that receive from recording/reproducing apparatus 12.In addition, according to the recommendation information that receives from program commending treatment facility 10, television display equipment 11 shows programs recommended information, and the processing of channel is set automatically.In addition, television display equipment 11 offers program commending treatment facility 10 to Operation Log.

On the other hand, according to the operation that the user carries out, recording/reproducing apparatus 12 records perhaps carry out recording reservation from the broadcast singal that TV receiving apparatus 4 receives, and according to the recommendation information that receives from program commending treatment facility 10, recording/reproducing apparatus 12 is the record recommend programs automatically.In addition, the recording medium reproducing program of recording/reproducing apparatus 12 from installing or embed, and on television display equipment 11 program of display reproduction.In addition, recording/reproducing apparatus 12 offers program commending treatment facility 10 to Operation Log.

Figure 31 is the block diagram of the structure of expression TV receiving apparatus 4.In the following description, TV receiving apparatus 4 is interpreted into the common receiving equipment of the standard criterion that meets digital broadcasting receiving apparatus.

Satellite ripple test section 91 is the signals according to the selection channel that receives from television display equipment 11 or recording/reproducing apparatus 12, detects and receive the assembly of satellite ripple broadcasting station 2 emissions via satellite and that received by antenna 3 selectively.Satellite ripple test section 91 also offers TMCC (emission and multiplexed configuration control) decoded portion 92 to the emission of explanation emission mode and multiplexed control signal, and the broadcast singal that the satellite ripple is held is offered demodulate/decode processing section 93.

TMCC decoded portion 92 is to receive the information be included in emission and the multiplexed control signal, and to the assembly of described information decoding.This information comprises the emission mode of time slot and indication modulator approach and encoding rate.TMCC decoded portion 92 offers demodulate/decode processing section 93 to decoded information.

Demodulate/decode processing section 93 is the methods of advising from the information about emission mode of TMCC decoded portion 92 receptions by adopting, to the assembly of the 91 broadcast singal demodulation sign indicating numbers that receive from satellite ripple test section.Demodulate/decode processing section 93 offers release of an interleave part 94 to demodulation sign indicating number result.The example of described method is QPSK (quarternary phase-shift keying (QPSK)) method and 8 PSK (phase shift keying) methods mutually.The QPSK method is also referred to as 4 phase modulator approaches or 4 phase PSK methods.

Release of an interleave part 94 is to the input signal release of an interleave, and the result that release of an interleave is handled is offered the assembly of correction process part 95.In addition, release of an interleave part 94 is also carried out frame separating treatment and scramble process to the release of an interleave result.

Correction process part 95 is input signal generally to be adopted the correction process of Reed-Solomon (Reed-Solomon) method, and the result of correction process is offered the assembly of CA (conditional access) descrambling part 101.

Earthwave test section 96 is the signal of basis from the selection channel of television display equipment 11 or recording/reproducing apparatus 12 receptions, the assembly of the earthwave of detection and reception antenna seizure selectively.Earthwave test section 96 also offers TMCC (emission and multiplexed configuration control) decoded portion 97 to the emission of explanation emission mode and multiplexed control signal, and the broadcast singal that earthwave is held is offered demodulate/decode processing section 98.

TMCC decoded portion 97 is to receive the information be included in emission and the multiplexed control signal, and to the assembly of described information decoding.This information comprises time stamp TS, the emission mode of time slot and indication modulator approach and encoding rate.TMCC decoded portion 97 offers demodulate/decode processing section 98 to decoded information.

Demodulate/decode processing section 98 is the methods of advising from the information about emission mode of TMCC decoded portion 97 receptions by adopting, to the assembly of the 96 broadcast singal demodulation sign indicating numbers that receive from the earthwave test section.Demodulate/decode processing section 98 offers release of an interleave part 99 to demodulation sign indicating number result.The example of described method is QAM (quadrature amplitude modulation) method.

Release of an interleave part 99 is to the input signal release of an interleave, and the result that release of an interleave is handled is offered the assembly of TS (transmitting stream) reproducing part 100.In addition, release of an interleave part 99 is also carried out frame separating treatment and scramble process to the release of an interleave result.

TS reproducing part 100 is to produce according to input signal to transmit stream, and transmitting the assembly that stream offers CA descrambling part 101.

CA descrambling part 101 is to the conditional access signal descrambling from correction process part 95 or 100 receptions of TS reproducing part, and the signal behind the descrambling is offered the assembly that multichannel is decomposed part 102.

Data input unit 103 is the assemblies that receive the stream data that EPG data and Distributor 5 from EPG receiving equipment 9 transmit by network 8.Data input unit 103 sends EPG data and stream data to multichannel decomposition part 102.

It is the signal from CA descrambling part 101 or data input unit 103 receptions to be carried out multichannel decompose that multichannel is decomposed part 102, thereby produce the audio signal that will be provided for audio signal decoding part 104, to be provided for the vision signal of vision signal decoded portion 105 and will be provided for the assembly of the data of data decoding portion 106.The data that offer data decoding portion 106 comprise control signal and EPG.

Audio signal decoding part 104 is the audio signal decodings to input, and the audio signal of decoding is offered the assembly of television display equipment 11 or recording/reproducing apparatus 12.By the same token, vision signal decoded portion 105 is the vision signal decodings to input, and the vision signal of decoding is offered the assembly of television display equipment 11 or recording/reproducing apparatus 12.Data decoding portion 106 is the data to input, for example the assembly of control signal and EPG decoding.Data decoding portion 106 offers television display equipment 11 or recording/reproducing apparatus 12 to decoded data.

In the processing that TV receiving apparatus 4 is carried out, by adopting preordering method to the earthwave of reception or the stream data demodulation sign indicating number of distribution, the result of demodulation sign indicating number is provided for television display equipment 11 or recording/reproducing apparatus 12.

Figure 32 is the block diagram of the structure of expression television display equipment 11.

Operation input section 121 is the operation inputs that receive from the user, and the signal that the operation that representative receives from the user is imported is offered the assembly of some assemblies that adopt television display equipment 11.In addition, operation input section 121 offers the Operation Log tabulation 122 that the user preserves described content to the content of the operation input that receives from the user.The Operation Log that adopts in the program commending treatment facility 10 of earlier in respect of figures 13 explanation obtains part 46 and reads the Operation Log that is kept in the Operation Log tabulation 122, the daily record of the operation of carrying out as the user.If the operation input that receives from the user is the operation of selecting channel, operation input section 121 offers channel to this operation input part 123 is set so.

It is signals of importing according to the operation of the representative of consumer input that receives from operation input section 121 that channel is provided with part 123, produces the assembly of the control signal of indication selected channel.Channel is provided with part 123 this control signal is offered TV receiving apparatus 4.In addition, according to the recommendation information that is kept in the programs recommended tabulation 128 that illustrates later, channel is provided with the control signal that part 123 produces the indication selected channel, and this control signal is offered TV receiving apparatus 4, is used for being provided with automatically the control signal of channel as TV receiving apparatus 4.TV receiving apparatus 4 receives the broadcast singal by the channel of described control signal regulation.

Data input unit 124 is from TV receiving apparatus 4 receiving broadcast signals, and this signal is sent to the assembly of image processing section 125.Image processing section 125 is the method for displaying image that adopt according to output 126, the broadcast singal of input is carried out the assembly of image processing.Image processing section 125 offers output 126 to the result of image processing.Output 126 comprises display unit and the audio output device such as loud speaker.The example of display unit is CRT (cathode ray tube) and LCD (LCD).Output 126 is the picture signals that show the input broadcast singal finish image processing, and from the assembly of audio output device output audio signal.

It is to obtain recommendation information from program commending treatment facility 10 that parts 127 are obtained in programs recommended tabulation, and the information that obtains is sent to the assembly of programs recommended tabulation 128.The recommendation information catalogue that 128 pairs of programs recommended tabulations provide.Part 123 is set channel and programs recommended information display control section 129 is read recommendation information from programs recommended tabulation 128.

Programs recommended information display control section 129 is that the recommendation information of reading from programs recommended tabulation 128 is offered image processing section 125, so that be used to show to the user assembly of programs recommended information.Image processing section 125 as independent information, perhaps as being superimposed upon by the information on the image of the broadcast singal representative that receives from data input unit 124, sends the recommendation information that receives from programs recommended information display control section 129 to output 126.Output 126 shows recommendation information on display unit.

Be noted that to allow the user, determine whether programs recommended information will be displayed on the output 126, and perhaps whether channel will be provided with automatically according to the recommendation information that receives from program commending treatment facility 10.

With reference to the flow chart shown in Figure 33, the recommendation information that the following basis that interpretation carried out receives from program commending treatment facility 10 shows that to the user recommendation information of programs recommended information shows processing.

At first, at step S251, the recommendation information that part 127 obtains recommendation information output 49 outputs of employing in program commending treatment facility 10 is obtained in programs recommended tabulation.

Subsequently, at next step S252, part 127 recommendation information catalogue to obtaining in programs recommended tabulation 128 is obtained in programs recommended tabulation.

Subsequently, at next step S253, programs recommended information display control section 129 is read the programs recommended information of broadcasting in the predetermined a period of time with respect to the current time from programs recommended tabulation 128, thereby produces the data of the information be used to show recommended program.The length of described a period of time was generally 3 hours or one day.The information of recommended program comprises the title of each recommended program, content, airtime and broadcasting channel.Programs recommended information display control section 129 offers image processing section 125 to these data subsequently.

Subsequently, at next step S254, image processing section 125 is carried out image processing, so that show the input data that are used for showing at output 126 information of recommended program, and the result of image processing is exported to output 126.The recommendation information of data representative can independent information or is superimposed upon form experience image processing by the information on the image of the broadcast singal representative that receives from data input unit 124.

Subsequently, at next step S255, output 126 shows the recommendation information that receives from image processing section 125.At last, finish the execution of the processing of this flow chart representative.

By carrying out above-mentioned processing, recommendation information is displayed on the output 126.Thereby by the recommendation information with reference to demonstration, the user can select him to want the program of watching.

With reference to the flow chart shown in Figure 34, the following basis that interpretation carried out is provided with the automatic channel set handling of the channel of broadcasting the program that conforms to user's hobby automatically from the recommendation information that program commending treatment facility 10 receives.

The processing of carrying out at step S271 and S272 is with identical with reference to the step S251 of the flow chart of Figure 33 explanation and processing that S252 carries out in front respectively.That is, the recommendation information of part 127 acquisitions by recommendation information output 49 outputs of adopting in the program commending treatment facility 10 obtained in programs recommended tabulation.Subsequently, the recommendation information catalogue of part 127 in 128 pairs of acquisitions of programs recommended tabulation obtained in programs recommended tabulation.

Subsequently, at next step S273, channel is provided with part 123 obtains the current time from program commending tabulation 128 programs recommended information.Subsequently, at next step S274, according to the swindleness of recommended program, channel is provided with part 123 and produces the channel configuration information, and the channel configuration information is exported to TV receiving apparatus 4 as control signal.According to this control signal, the broadcast singal that TV receiving apparatus 4 receives by the channel broadcasting of stipulating in the control signal.

Subsequently, at next step S275, data input unit 124 obtains the broadcast singal of specified channels from TV receiving apparatus 4, and this signal is offered image processing section 125.

Subsequently, at next step S276, the broadcast singal of 125 pairs of inputs of image processing section carries out image processing, so that display image, and result is offered output 126.

Subsequently, at next step S277, output 126 is with the form of the image that obtains as the result of image processing, shows the image of the programs recommended information that receives from image processing section 125, and output sound.At last, finish the execution of the processing of this flow chart representative.

By carrying out above-mentioned processing, the channel of the program that broadcasting conforms to user's hobby is provided with automatically.For example when the user sends order about this processing, above can carrying out with reference to the automatic channel set handling of the flowchart text shown in Figure 34.Thereby when the user exists when determining to watch the problem of which program, the channel of broadcasting appropriate program can be provided with automatically.

In addition, when each predetermined a period of time that the user continues a period of time of being considered to ignore is not carried out input operation, perhaps predetermined a period of time past tense of a period of time that ought be considered to ignore, with reference to the automatic channel set handling of the flowchart text shown in Figure 34, described predetermined a period of time was generally 2 hours above can carrying out.Predetermined a period of time of a period of time that is considered to ignore, to be that the user is clear therebetween knew that the user is not provided with a period of time of the fact of channel.

Be noted that also can allow the user to import setting can not carry out the order of the pattern of automatic channel set handling.By such forbidding automatic channel set handling, when the user had a mind to watch program by a certain channel broadcasting, this channel can not be changed over another channel automatically.

Figure 35 is the block diagram of expression recording/reproducing apparatus 12.

Operation input section 141 is the operation inputs that receive from the user, and the control signal that the operation that representative receives from the user is imported is offered the assembly of other assembly that constitutes recording/reproducing apparatus 12.In addition, operation input section 141 offers the content of the operation input that receives from the user Operation Log tabulation 142 that is used to preserve described content.The Operation Log that adopts in the program commending treatment facility 10 of earlier in respect of figures 13 explanation obtains part 46 and reads the Operation Log that is kept in the Operation Log tabulation 142, the daily record of the operation of carrying out as the user.

Recording setting part 143 is according to receiving from operation input section 141, the signal of the operation input of representative of consumer input, the recording processing information necessary is carried out in extraction, perhaps extracts the assembly that carries out the recording processing information necessary from the recommendation information that is kept in the programs recommended tabulation 149 that illustrates later.Carry out the recording processing information necessary and comprise the broadcasting time started of the program that will write down and the channel of off-the-air time and broadcast program.If is the operation of carrying out recording reservation from the control signal indication of operation input section 141 receptions from the operation input that the user receives, recording setting part 143 is made a catalogue to carrying out the recording processing information necessary in recording reservation tabulation 144 so.If the operation input that the control signal indication that receives from operation input section 141 receives from the user is the request of the program of record current broadcast, perhaps in the time will being kept at recommendation information in the programs recommended tabulation 149 that illustrates later by utilization and writing down automatically, recording setting part 143 offers record controls part 145 carrying out the recording processing information necessary.

Record controls part 145 is according to what receive from recording setting part 143 program to be carried out the recording processing information necessary, produce the control signal that the broadcasting channel of the program that will write down is broadcasted in indication, perhaps, produce the assembly of control signal by recording reservation information from the 144 recording reservation information extraction current time of making a catalogue of tabulating at recording reservation.Record controls part 145 offers TV receiving apparatus 4 to control signal.Record controls part 145 still produces the control signal that writes down, and this control signal is offered the assembly of recoding/reproduction processing section 147.According to the control signal that receives from record controls part 14, the broadcast singal that TV receiving apparatus 4 receives by the channel broadcasting of regulation.

Data input unit 146 is from TV receiving apparatus 4 receiving broadcast signals, and this signal is sent to the assembly of recoding/reproduction processing section 147.Recoding/reproduction processing section 147 generally has and allows dismountable recording medium to be installed in structure on the recoding/reproduction processing section 147, perhaps comprises the inner recording medium that embeds.The example of dismountable recording medium is a tape, CD, and disk, magneto optical disk and semiconductor memory, and the inner recording medium that embeds is hard disk and semiconductor memory.Recoding/reproduction processing section 147 can record information on the recording medium, and from recording medium reproducing information dismountable or that embed.More particularly, suppose to be tape according to the detachable recording medium on record recoding/reproduction processing section 147.In this case, recoding/reproduction processing section 147 is furnished with in recording processing, the broadcast singal that receives from data input unit 146 is recorded on the tape and from the magnetic head of tape information reproduction.147 information of reproducing in recoding/reproduction processing section offer the equipment such as television display equipment 11, so that show described information.

It is to obtain recommendation information from program commending treatment facility 10 that part 148 is obtained in programs recommended tabulation, and this information is passed to the assembly of programs recommended tabulation 149.Programs recommended tabulation 149 is the assemblies that are used for the recommendation information catalogue that provides.Recording setting part 143 is read and will be used for self registering recommendation information from programs recommended tabulation 149.

Below with reference to the automatic record of the flowchart text shown in Figure 36.

At first, at step S291, programs recommended tabulation is obtained part 148 and obtain recommendation information from the recommendation information output 49 that adopts program commending treatment facility 10.

Subsequently, at next step S292, the recommendation information catalogue of part 148 in 149 pairs of acquisitions of programs recommended tabulation obtained in programs recommended tabulation.

Subsequently, at next step S293, recording setting part 143 extracts the programs recommended information of current time from programs recommended tabulation 149.Recording setting part 143 is the required information of information acquisition record from extracting subsequently, and acquired information is offered record controls part 145.The record information needed that obtains comprises broadcasting time started and the off-the-air time of wanting recorded program and the channel of broadcasting this channel.

Subsequently, at next step S294, record controls part 145 produces the channel configuration information of the broadcast singal that is used to receive the program that will write down, and the channel configuration information is exported to TV receiving apparatus 4 as control signal.According to this control signal, TV receiving apparatus 4 receives the broadcast singal by the program of the channel broadcasting of regulation.

Subsequently, at next step S295, data input unit 146 obtains the broadcast singal of specified channels from TV receiving apparatus 4, and this signal is sent to recoding/reproduction processing section 147.

Subsequently, at next step 296, recoding/reproduction processing section 147 records the broadcast singal that receives from TV receiving apparatus 4 on the recording medium that embeds installation or inner.At last, finish the execution of the processing of this flow chart representative.

By carrying out above-mentioned processing, can write down the program that conforms to user's hobby automatically.When carrying out the recording processing of user's request, the recording processing that has perhaps begun when for example Yu Yue recording processing is being carried out, is not carried out top automatic recording processing with reference to the flowchart text shown in Figure 36.

Above interpretation according to the automatic recording processing of carrying out for the current time recommend programs.But needless to say, also can carry out automatic recording processing according to identical mode according to being the recommendation information that falls behind a certain moment acquisition of predetermined a period of time than the current time.Like this, can realize the reservation of writing down automatically.

In each processing with reference to figure 1-36 explanation, program vector PP produces in Distributor 5 in front.But, replacing in Distributor 5, producing program vector PP, Distributor 5 can offer the program commending treatment facility to the EPG data by network 8, so that be used to produce program vector PP in the program commending treatment facility.

Figure 37 is about not being to produce program vector PP in Distributor 171, but according to passing through network 8, offer the EPG data of program commending treatment facility 191 from Distributor 171, in program commending treatment facility 191, produce the situation of program vector, the block diagram of the structure of expression Distributor 171.In this case, the structure that also is used for producing the program commending treatment facility 191 of program vector PP is shown in Figure 38.

Be noted that, the part identical with the corresponding counterpart that adopts in the program commending treatment facility 10 shown in the Distributor shown in Fig. 25 and Figure 13 represented with the Reference numeral identical with their respective counter, for fear of repetition, no longer repeat their explanation.

That is, Distributor 171 comprises such as described in reference to Figure 2, is used in data acquiring portion 21 and tcp data segment 25 in the Distributor 5 equally.Distributor 171 is read stream data from stream data database 6, perhaps from metadata database 7 or comprise that the EPG data of metadata read metadata, and by network 8 stream data or metadata is sent to EPG receiving equipment 9 or TV receiving apparatus 4.

Program commending treatment facility 191 has the structure identical with the program commending treatment facility 10 of earlier in respect of figures 13, and the metadata that adopts in the Distributor 5 shown in Fig. 2 extracts part 22 and program vector produces the part 23 except program commending treatment facility 191 also is included in.Except the processing that program commending treatment facility 10 is carried out, the program vector that program commending treatment facility 191 also carries out the flowchart text shown in the earlier in respect of figures 3 produces processing 1, the program vector of the flowchart text shown in the earlier in respect of figures 6 produces handles 2, the packet transaction 1 of the flowchart text shown in the earlier in respect of figures 7 and the packet transaction 2 of the flowchart text shown in the earlier in respect of figures 8.

Replace producing program vector PP, Distributor 171 can offer program commending treatment facility 191 to the EPG data by network 8, for use in producing program vector PP.Although there is this difference, still can obtain identical result with earlier in respect of figures 1-36 explanation.

In addition, also can realize a kind of like this system configuration, wherein EPG receiving equipment 9 is collected the information and the configuration information of the operation that users carry out from television display equipment 11 and recording/reproducing apparatus 12, and an information of collecting offers Distributor 201, and Distributor 201 not only produces program vector PP, and carry out matching treatment, by network 8 result of matching treatment is offered EPG receiving equipment 9.With regard to this system configuration, network has the structure shown in Figure 39, and Distributor 201 has the block diagram shown in Figure 40.

Be noted that the system configuration assembly that is equal to their respective counter shown in Fig. 1,37 and 38 represented by the Reference numeral identical with their respective counter, and, omitted explanation them for fear of repetition.

That is, except the function of the Distributor 171 of earlier in respect of figures 37 explanation, Distributor 201 also has the function of the program commending treatment facility 191 of earlier in respect of figures 38 explanations.Thereby the user does not need to have the program commending treatment facility

In the structure shown in Figure 39 and 40, Distributor 201 carries out all every processing of earlier in respect of figures 3-12 and Figure 14-30 explanation.These processing comprise and produce vector, program vector PP for example, program side effect vector EfUP, positive history vectors UP, negative history vectors MUP and standard are liked the processing of vectorial APP, to the processing of program vector PP grouping, matching treatment and select unusual programs recommended processing.

Be noted that this moment, the history of the operation that the user carries out and configuration information generally have attached to it, are used to discern because of the different history of user and the user ID of configuration information.EPG receiving equipment 9 is collected operation history and configuration information from television display equipment 11 and recording/reproducing apparatus 12, by network 8 described history and information is sent to Distributor 201.The program commending treatment facility 191 that adopts in Distributor 201 is by according to the user ID that is attached on described operation history and the configuration information, to described operation history and configuration information classification, the operation history and the configuration information that receive from EPG receiving equipment 9 are kept at such as initial directory stores part 45, in the sector of breakdown of positive historical storage part 47 and negative historical storage part 48.

Above-mentioned interpretation wherein by using the EPG data of television broadcasting signal, recommend the situation of the program that conforms to user's hobby.But the present invention also can be applicable to wherein by attribute information being added in radio broadcasting and the various digital contents such as stream data the situation of the program that recommendation conforms to user's hobby.

The processing of each series that illustrates previously also can realize by the execution of software.If realize the processing of each series described above by the execution of software, the program that constitutes described software so can generally be installed to the computer that embeds the specialized hardware from network or program recorded medium, in the general purpose personal computer etc.By various programs are installed in the general purpose personal computer, personal computer can be realized various functions.

The program recorded medium that above mentioned handle the program in computer or the general purpose personal computer of will being installed to is recorded as the program that will be carried out by described computer or general purpose personal computer respectively is the bag medium that is independent of computer or general purpose personal computer distribution and offers the user.As preceding with reference to as described in figure 2,13 or 38, the example of bag medium comprises the

disk

31 or 71 such as floppy disk,

CD

32 or 72 such as CD-ROM (Compact Disc-Read Only Memory) or DVD (digital universal disc), magneto

optical disk

33 or 73 such as MD (MiniDisc) (trade mark), and

semiconductor memory

34 or 74.

In addition, in this manual, be kept at the step of each program in the recording medium according to predefined procedure along time shaft.But, needn't carry out described step according to predefined procedure along time shaft.For example, also can be simultaneously or be kept at the step of each program in the recording medium individually.

Be noted that the technical term " system " that uses in this specification means the fusion structure that comprises a plurality of equipment.

Industrial Applicability A

As mentioned above, according to the present invention, can produce the attribute information of content. Especially, can The attribute information of content is associated with weighted value, and each weighted value is stipulated every in a plurality of projects Contribution journey to the calculating of the similarity between attribute information and predesignated subscriber's the hobby information Degree.

In addition, according to another invention, not only can select the content that conforms to user's hobby, But also can be by using in a plurality of projects of regulation each to attribute information and predesignated subscriber Hobby information between the weighted value information of percentage contribution of calculating of similarity, in selecting Hold. Thereby, can select the content that correctly conforms to user's hobby.

In addition, according to another invention, can find out user's hobby. Especially, by comparing User's hobby and the hobby of general public can be identified the deviation (bias) of user's hobby, Thereby can determine the hobby that the user is exclusive.

Claims

1, a kind of messaging device of carrying out the processing of chosen content comprises:

Obtain the deriving means of the attribute information that comprises a plurality of projects of described content; With

Utilize predetermined weighted value information calculations by the described attribute information of described deriving means acquisition and from the calculation of similarity degree device between the information of user's acquisition,

Wherein said weighted value information specifies is according to described attribute information with comprise the described information that obtains from the user of at least a portion in the middle of the included a plurality of described project of described attribute information, calculating is as the project similarity of the similarity of each described project, and when calculating described attribute information and described similarity between the information that the user obtains, corresponding to the described project similarity percentage contribution separately of a plurality of described projects according to described project similarity.

2, according to the described messaging device of claim 1, wherein except the described attribute information of described content, described deriving means also obtains described weighted value information,

Described calculation element is by utilizing described weighted value information, and the attribute information of more described content and the described information that obtains from the user are calculated described similarity.

3, according to the described messaging device of claim 1, wherein said weighted value information is to disclose the information of the project that described user is overstated in the described attribute information of described content want.

4, according to the described messaging device of claim 1, also comprise:

Receive the input device of described user's operation input,

The utilization of wherein said weighted value information is set by the described user's of described input device input operation input.

5, according to the described messaging device of claim 1, also comprise:

Preserve storage device with the described weighted value information of described corresponding items of the described attribute information that obtains by described deriving means.

6, according to the described messaging device of claim 5, wherein said calculation element is from the described weighted value information that described storage device is preserved, extraction meets the described weighted value information of the condition that described content has, and utilizes the described weighted value information of being extracted out to calculate described attribute information and described similarity between the information of user's acquisition that described deriving means obtains.

7, a kind of information processing method of messaging device of the processing of carrying out chosen content comprises:

Obtain the obtaining step of the attribute information that comprises a plurality of projects of described content; With

Utilize predetermined weighted value information calculations by the described attribute information of the processing acquisition of described obtaining step and from the calculation of similarity degree step between the information of user's acquisition,

8, according to the described information processing method of claim 7, wherein in the processing of described obtaining step, except the described attribute information of described content, also obtain described weighted value information,

In the processing of described calculation procedure, by utilizing described weighted value information, the attribute information of more described content and the described information that obtains from the user are calculated described similarity.

9, according to the described information processing method of claim 7, wherein said weighted value information is to disclose the information of the project that described user is overstated in the described attribute information of described content want.

10, according to the described information processing method of claim 7, also comprise:

Receive the operation input step of described user's operation input,

The utilization of wherein said weighted value information is set by the described user's of the processing input of described operation input step operation input.

11, according to the described information processing method of claim 7, also comprise:

Preserve storing step with the described weighted value information of described corresponding items of the described attribute information of processing acquisition by described obtaining step.

12, according to the described information processing method of claim 11, wherein in the processing of described calculation procedure, from the described weighted value information of preserving by the processing of described storing step, extraction meets the described weighted value information of the condition that described content has, and utilizes the described weighted value information of being extracted out to calculate described attribute information and described similarity between the information of user's acquisition that the processing by described obtaining step obtains.

13, a kind of program that is used to make computer to carry out the processing of chosen content comprises:

14, a kind of recording medium has write down the described program of claim 13.

15, a kind of messaging device of carrying out the processing of chosen content comprises:

Preserve first storage device of the attribute information that comprises a plurality of projects of described content; With

Utilize predetermined weighted value information calculations by the described attribute information of described first storage device preservation and from the calculation of similarity degree device between the information of user's acquisition,

16, according to the described messaging device of claim 15, wherein except the described attribute information of described content, described first storage device is also preserved described weighted value information,

17, according to the described messaging device of claim 15, wherein said weighted value information is to disclose the information of the project that described user is overstated in the described attribute information of described content want.

18, according to the described messaging device of claim 15, also comprise:

Receive the input device of described user's operation input,

19, according to the described messaging device of claim 15, also comprise:

Preserve second storage device with the described corresponding items weighted value information of the described attribute information of preserving by described first storage device.

20, according to the described messaging device of claim 19, wherein said calculation element is from the described weighted value information that described second storage device is preserved, extraction meets the described weighted value information of the condition that described content has, and utilizes the described weighted value information of being extracted out to calculate described attribute information and described similarity between the information of user's acquisition that described first storage device is preserved.

21, a kind of information processing method of carrying out the processing of chosen content comprises:

Preserve first storing step of the attribute information that comprises a plurality of projects of described content; With

Utilize predetermined weighted value information calculations by the described attribute information of the processing preservation of described first storing step and from the calculation of similarity degree step between the information of user's acquisition,

22, according to the described information processing method of claim 21, wherein in the processing of described first storing step, except the described attribute information of described content, also preserve described weighted value information,

23, according to the described information processing method of claim 21, wherein said weighted value information is to disclose the information of the project that described user is overstated in the described attribute information of described content want.

24, according to the described information processing method of claim 21, also comprise:

Receive the operation input step of described user's operation input,

25, according to the described information processing method of claim 21, also comprise:

Preserve second storing step with the described corresponding items weighted value information of the described attribute information of processing preservation by described first storing step.

26, according to the described information processing method of claim 25, wherein in the processing of described calculation procedure, from the described weighted value information of preserving by the processing of described second storing step, extraction meets the described weighted value information of the condition that described content has, and the described attribute information and described similarity between the information of user's acquisition that utilize the described weighted value information of being extracted out to calculate to preserve by the processing of described first storing step.

27, a kind of program that is used to make computer to carry out the processing of chosen content comprises:

28, a kind of recording medium has write down the described program of claim 27.