JP4275635B2 - Preference level calculation device, content recommendation device, content record output device, preference level calculation server, preference level calculation method, preference level calculation program, and computer-readable recording medium recording the same - Google Patents

Description

  The present invention relates to a preference level calculation device, a content recommendation device, a content recording output device, a preference level calculation server, a preference level calculation method, a preference level calculation program, and a preference level calculation program for calculating a preference level for a keyword related to a content being viewed. The present invention relates to a computer-readable recording medium.

  In recent years, there has been proposed a device that uses an electronic program guide to search for and recommend a television program suitable for a user and automatically record the television program. In these devices, in order to collect user preference information, a keyword is extracted from electronic program guide information of a program viewed by the user or a program recorded by the user, and the extracted keyword is stored as the preference information of the user. is doing.

  For example, in Patent Document 1, keywords appearing in a program description in an electronic program guide of a program selected and recorded by a viewer are extracted, and the frequency of appearance of the keywords is counted and stored. A video device that recommends a program having a high appearance frequency of the keyword when making a recording reservation is disclosed.

In Patent Document 2, electronic program guide information of a program viewed by a user is disassembled, keywords obtained thereby are accumulated, parameters such as a keyword type such as a program title and performers, and a keyword viewing frequency are stored. A program recommendation device is disclosed that calculates a preference point that represents a user's preference degree with respect to a keyword and recommends a program including the keyword based on the calculated preference point.
JP 7-135621 A (published May 23, 1995) JP 2002-142160 A (released February 17, 2002)

  However, in the video equipment disclosed in Patent Document 1, electronic program guide information of a program preferred by the user is decomposed into keywords by morphological analysis or the like, and the keywords are only accumulated. Specifically, in order to count the appearance frequency of keywords, points are added uniformly to all extracted keywords, the amount of points is adjusted according to the viewing time of the user's viewing program, and the viewer records The keywords extracted from the program are counted and accumulated by increasing the weight and adding points.

  On the other hand, in the program recommendation device disclosed in Patent Document 2, keywords are accumulated for each profile by dividing into a plurality of profiles such as program title, performer, or genre.

  Here, in evaluating the degree of preference for a keyword related to a program viewed by the user, not only information related to the program viewed by the user but also information not directly related to the program viewed by the user is effective. For example, information such as the contents of the back program broadcast behind the program viewed by the user and the fashion at the time the program was broadcast is also effective.

  Specifically, when the viewer views the content, for example, the weight of the viewer viewing the content differs depending on the content of the back program during the time zone when the content was broadcast. That is, when another content is viewed despite the fact that the back program with a high degree of preference is being broadcast, it can be determined that the viewer's degree of preference for the content is very high and, as a result, extracted from the content. It can be determined that the keyword reflects the viewer's preference very well.

  However, in the devices disclosed in Patent Literature 1 and Patent Literature 2, keywords are not accumulated in consideration of such information. For this reason, it is not possible to appropriately calculate the degree of preference for keywords related to programs viewed by the user.

  Accordingly, an object of the present invention is to provide a preference level calculation device, a content recommendation device, a content recording output device, a preference level calculation server, and a preference level that can appropriately calculate a preference level for a keyword related to content being viewed by a user. The object is to provide a calculation method, a preference calculation program, and a computer-readable recording medium on which the program is recorded.

  The preference calculation device of the present invention is a keyword extraction means for extracting a keyword from content information indicating information relating to the first content, and the first content distributed when the first content and the first content are viewed. The preference value for the keyword extracted by the keyword extraction means based on the evaluation value calculation means for calculating the evaluation value for the preference for the second content, and the evaluation value for the first content and the evaluation value for the second content Keyword preference degree calculating means for calculating the degree.

  The preference extraction method of the present invention includes a keyword extraction step of extracting a keyword from content information indicating information related to the first content, and the second content distributed when the first content and the first content are viewed. Degree of preference for the keyword extracted in the keyword extraction step based on the evaluation value calculation step for calculating the evaluation value related to the preference of the content, and the evaluation value of the first content and the evaluation value of the second content And a keyword preference degree calculating step for calculating.

  According to the above configuration, the keyword extracting means (keyword extracting step) extracts a keyword from the content information indicating the information related to the first content that is the content viewed by the user (hereinafter referred to as viewing content), and the viewing content. The keyword preference degree calculating means (keyword) for the extracted keyword based on the evaluation value regarding the preference for the first content and the second content distributed when the user is viewing the viewing content. (Preference level calculation step) calculates the preference level.

  Thus, instead of giving a uniform preference level to the keywords extracted from the viewing content, the evaluation values related to the preferences for the first content and the second content are used to The degree of preference can be calculated appropriately. For example, by setting the evaluation value related to the preference for the first content and the second content as the user's preference level for each content, the keyword is based on the preference level of the first content and the preference level of the second content. The preference level calculation means (keyword preference level calculation step) can calculate the preference level for the keyword.

  In the above preference degree calculating device, preferably, the content information is an electronic program guide indicating the contents of the content to be distributed.

  According to said structure, the keyword regarding the said viewing content is extracted from the program information regarding the viewing content in the electronic program guide delivered from the internet or a broadcast station.

  Here, the electronic program guide is information indicating the content of content distributed (broadcast) from the Internet or a broadcasting station, and includes program information of a television program, for example. The electronic program guide includes information such as an ID for identifying a program, broadcast start date / time, broadcast end date / time, broadcast channel, genre, sub-genre, program name, and program content.

  Thus, if the preference calculation device is in an environment where the electronic program guide can be used, the program information related to the viewing content viewed by the user is searched from the electronic program guide, and the keyword related to the viewing content is extracted from the program information. be able to. Then, the keyword preference level calculation means can appropriately calculate the preference level for the extracted keyword based on the evaluation value regarding the preference for the first content and the second content as the viewing content. .

  In the above preference degree calculating device, preferably, the first content holds content meta information as information related to the content, and the content information is the content meta information.

  According to the above configuration, keywords are extracted from the content meta information held by the first content that is the viewing content.

  Here, the content meta information is information related to the content given to the content. For example, information such as a content name, a broadcast start date / time, a broadcast end date / time, a broadcast channel, performer information, a genre, and a keyword is included. include.

  Thus, if content meta information is given to the viewing content viewed by the user, a keyword related to the viewing content can be extracted from the content meta information given to the viewing content. Then, the keyword preference level calculation means can appropriately calculate the preference level for the extracted keyword based on the evaluation value regarding the preference for the first content and the second content as the viewing content. .

  In the above-described preference level calculation device, preferably, a character data recognition unit that recognizes character data in the video information of the first content is provided, and the content information is character data obtained from the character data recognition unit.

  According to the above configuration, the character data recognizing means recognizes character data such as telop included in the video information of the viewing content, converts it to text data, and extracts keywords related to the viewing content from the converted text data. To do.

  Thus, if there is character data such as a telop in the video information of the viewing content viewed by the user, the character data recognition means recognizes the character data and converts it into text data. Keywords related to viewing content can be extracted from the text data. Then, the keyword preference level calculation means can appropriately calculate the preference level for the extracted keyword based on the evaluation value regarding the preference for the first content and the second content as the viewing content. .

  In the above preference level calculation device, preferably, the apparatus includes a voice data recognition unit that recognizes voice data included in the first content and converts it into text data, and the content information is text data obtained from the voice data recognition unit. is there.

  According to said structure, audio | voice data recognition means recognizes audio | voice data, such as a performer's audio | voice included in viewing content, converts it into text data, and extracts the keyword regarding viewing content from the converted text data. .

  Thus, if the audio content such as the performer's voice is present in the viewing content viewed by the user, the audio data recognition means recognizes the audio data and converts it into text data. Keywords related to viewing content can be extracted from text data. Then, the keyword preference level calculation means can appropriately calculate the preference level for the extracted keyword based on the evaluation value regarding the preference for the first content and the second content as the viewing content. .

  In the above preference level calculation device, preferably, the preference level calculation unit includes a preliminary preference level calculation unit that calculates the user's preference level for each of the first content and the second content, and the evaluation value regarding the preference for the content is the preliminary preference level. The preference degree for each content obtained from the calculation means, and the keyword preference degree calculation means includes a preference degree of the first content and a preference degree of the second content for which the preference degree is calculated by the preliminary preference degree calculation means. Based on the above, the degree of preference for the keyword related to the first content is calculated.

  According to the above configuration, the preliminary preference level calculation unit calculates the user's preference level for each of the first content and the second content, and the calculated preference level of each content is an evaluation value related to the preference for the content. The keyword preference level calculation means calculates the preference level for the keyword extracted from the viewing content.

  Here, the content preference level is information indicating how much the content matches the user's preference. This is because the user's preference for the content is calculated using information representing the user's preference, such as the user's favorite person, favorite genre, favorite program name, or favorite keyword, and the degree of preference is high. As shown, the content matches the user's preference. Note that the information indicating the user's preference may be input by the user himself / herself, or may be automatically extracted from the user's past viewing history, action history, or the like. Further, the user's preference level for the content may be calculated within the above-described preference level calculation device, or may be calculated outside the preference level calculation device and input to the preference level calculation device. It doesn't matter.

  Thereby, the keyword preference degree calculating means extracts the keyword extracted from the viewing content based on the user's preference degree with respect to the viewing content and, for example, the user's preference degree with respect to a plurality of contents broadcast simultaneously other than the viewing content. It is possible to appropriately calculate the degree of preference.

  In the above-described preference level calculation device, preferably, the keyword preference level calculation means is based on the preference level of the first content and the preference level of the second content having a higher preference level than the first content. The degree of preference for keywords related to one content is calculated.

  Thereby, when content having a higher preference level than the user's preference level for the viewing content is broadcast behind the scenes, for example, according to the difference in the preference level of the content, The degree of preference for the extracted keyword can be greatly calculated. In addition, the above-mentioned means that the viewing content was viewed without viewing the content with the high degree of preference, even though the content with the high degree of preference was broadcast behind the viewing content. In this case, the viewing content is content that matches the user's preference, and the keywords extracted from the viewing content can be regarded as well representing the user's preference.

  In the above preference level calculation device, preferably, for each of the first content and the second content, a trend level calculation unit that calculates a trend level for each content is provided, and the evaluation value regarding the preference for the content is the trend level The degree of popularity of each content obtained from the calculating means, and the keyword preference degree calculating means is configured to determine the popularity degree of the first content and the popularity degree of the second content for which the popularity degree is calculated by the popularity degree calculating means. Based on the first content, the degree of preference for the keyword is calculated.

  According to the above configuration, the trend level calculation means calculates the trend level for each of the first content and the second content, and uses the calculated trend level of each content as an evaluation value related to the preference for the content. The preference level calculation means calculates the preference level for the keyword extracted from the viewing content.

  Here, the degree of popularity is information indicating whether the content is content in line with the fashion of the world, or whether the content is content suitable for the current season. For example, content in which actors that are currently being talked about in the world, content that has attracted public attention and disasters, and content related to the Olympics at the time of the Olympic Games are equivalent to highly popular content. To do.

  Thus, the keyword preference level calculation means calculates the preference for the keyword extracted from the viewing content based on the trend level for the viewing content and the trend level for a plurality of contents broadcast simultaneously other than the viewing content, for example. The degree can be calculated appropriately.

  In the above preference level calculation device, preferably, the keyword preference level calculation means is configured to determine the first content based on the trend level of the first content and the trend level of the second content having a higher trend level than the first content. The degree of preference for keywords related to one content is calculated.

  As a result, when content having a higher fashion than the popularity of the viewing content is broadcast on the back, for example, the larger the difference is, the more the above-mentioned difference is extracted from the viewing content. It is possible to greatly calculate the degree of preference with respect to the keyword. In addition, the above-mentioned means that the viewing content was viewed without being caught by the fashion, even though the content with a high fashion level is broadcast behind the viewing content. In this case, the viewing content is content that matches the user's preference, and the keywords extracted from the viewing content can be regarded as well representing the user's preference.

  The preference level calculation device preferably includes audience rating input means for inputting an audience rating for each of the first content and the second content, and an evaluation value related to the preference for the content is obtained from the audience rating input means. The keyword preference level calculating means calculates the first audience rating based on the audience rating of the first content and the audience rating of the second content. The degree of preference for keywords related to one content is calculated.

  According to said structure, an audience rating input means inputs the audience rating about each of 1st content and 2nd content, and uses the audience rating of each input content as an evaluation value regarding the preference about content, The keyword preference level calculation means calculates the preference level for the keyword extracted from the viewing content.

  Here, the audience rating is information indicating how many people who can receive the content have viewed the content, and content with a high audience rate is the person who has viewed the content. It shows that there are many. Note that the audience rating is calculated by an external device and is input to the preference level calculation device via the audience rating input means.

  Thus, the keyword preference level calculation means calculates the preference for the keyword extracted from the viewing content based on the viewing rate for the viewing content and, for example, the viewing rate for a plurality of contents broadcast simultaneously other than the viewing content. The degree can be calculated appropriately.

  In the above-described preference level calculation device, preferably, the keyword preference level calculation unit is configured to change the first content based on the audience rating of the first content and the audience rating of the second content having a higher audience rating than the first content. The degree of preference for keywords related to one content is calculated.

  As a result, when content with a higher audience rating than the audience content is broadcast behind the scenes, for example, the larger the above difference, the more the above-mentioned difference is extracted from the audience content. It is possible to greatly calculate the degree of preference with respect to the keyword. In addition, the above means that the viewing content was viewed without viewing the content with the high audience rating, even though the content with the large audience rating was broadcast behind the viewing content. In this case, the viewing content is content that matches the user's preference, and the keywords extracted from the viewing content can be regarded as well representing the user's preference.

  A content recommendation device according to the present invention includes any one of the above-described preference level calculation devices, a keyword storage unit that stores the preference level of the keyword calculated by the keyword preference level calculation unit and the keyword, and a keyword storage. The apparatus further includes recommended content determining means for determining a user's preference level for content based on information stored in the means and determining recommended content for the user.

  According to the above configuration, the preference level for the keyword extracted from the viewing content is calculated by the preference level calculation device, the keyword storage unit stores the keyword and the preference level for the keyword in association with each other, and the recommended content determination unit The recommended content for recommending the content to the user is determined using the information.

  Thus, the preference level for the keyword extracted from the viewing content by the preference level calculation device can be stored in the keyword storage unit, and the preference level for the keyword can be stored by the keyword preference level calculation unit of the preference level calculation device. Each time is calculated, the user's preference information can be appropriately reflected by updating the information stored in the keyword storage means. Furthermore, by using the information stored in the keyword storage unit appropriately reflecting the user's preference information, it is possible to appropriately recommend content that matches the user's preference.

  The content recommendation device preferably further includes a display unit that displays the preference level of the keyword calculated by the keyword preference level calculation unit.

  According to said structure, the preference degree about the keyword calculated by the preference degree calculation apparatus is displayed on a display apparatus, for example.

  As a result, the keyword extracted from the viewing content and the preference level of the keyword are displayed on, for example, a display device, so that the user can grasp how the user's own preference information is reflected in the keyword storage means. can do.

  A content recording / output device according to the present invention includes the preference calculation device according to any one of the above, and further includes a content recording unit that records content and a content output unit that outputs the content recorded by the content recording unit. I have.

  According to the above configuration, the content recording unit records the content, and the content output unit outputs the content recorded by the content recording unit to, for example, the display device and displays it.

  Thereby, even when the user reproduces and views the recorded content, the output is performed based on, for example, the user's preference for a plurality of contents broadcast when the viewing content is output and reproduced. It is possible to appropriately calculate the degree of preference for the keyword extracted from the content being reproduced.

  A preference level calculation server of the present invention includes the preference level calculation device according to any one of the above, and further includes viewing history information input means for inputting viewing history information by a user, and the keyword extraction means includes viewing history information input means. The content included in the viewing history information input by is regarded as the first content.

  According to said structure, a preference degree calculation server inputs the viewing history information which the user watched the content via viewing history information input means. Then, the keyword extracting means regards the content included in the input viewing history information as the first content, so that the keyword extracting means extracts the keyword from the content information indicating the information related to the first content, and the keyword preference level. The calculating means can appropriately calculate the degree of preference for the keyword. The viewing history information input means inputs viewing history information via wired or wireless.

  As a result, it is possible to achieve the same effect by performing the same processing as described above in the preference level calculation server, and further, each processing is performed on the server side. Can be reduced.

  The preference calculation device of the present invention is a keyword extraction means for extracting a keyword from content information indicating information relating to the first content, and the first content distributed when the first content and the first content are viewed. The preference value for the keyword extracted by the keyword extraction means based on the evaluation value calculation means for calculating the evaluation value for the preference for the second content, and the evaluation value for the first content and the evaluation value for the second content And a keyword preference degree calculating means for calculating the degree.

  As a result, it is suitable for each keyword by using the evaluation value related to the preference for the first content and the second content, instead of uniformly giving preference to the keywords extracted from the viewing content. The degree of preference can be calculated.

  Furthermore, the evaluation value related to the preference for the first content and the second content is information such as the user's preference level for the content, the fashion level for the content, and the audience rating for the content. Therefore, it is possible to calculate a more suitable degree of preference for the keyword extracted from the viewing content.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the drawings used for the following description, the same parts are denoted by the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated.

(Functional configuration of preference level calculation device)
FIG. 1 shows a block diagram of a functional configuration of the preference degree calculation apparatus 100 according to the present embodiment. As illustrated in FIG. 1, the preference level calculation device 100 includes an input unit 101, a content input processing unit 102, and an audience rating input unit 103. Furthermore, the preference level calculation device 100 includes a trend level calculation unit 104, a preliminary preference level calculation unit 105, a character data recognition unit 106, a voice data recognition unit 107, a keyword extraction unit 108, a keyword preference level calculation unit 109, and an output processing unit. 110 is provided.

  The content 10 includes not only video media recorded on, for example, a television program or recording media broadcasted wirelessly or via wire, but also music media such as radio programs and music provided in the same manner. That is, it refers to all multimedia contents provided by various forms and various media.

  The input unit 101 is an interface for receiving an electronic program guide distributed via, for example, wireless or wired, and is used to input the electronic program guide 12 via a network.

  The electronic program guide 12 may be distributed via the Internet, may be distributed from a broadcasting station, or may read information recorded on a recording medium or the like. Absent. Further, the electronic program guide 12 may be automatically input to the preference level calculation device 100 or may be input by a viewer's operation.

  The content input processing unit 102 is an interface that receives content broadcasted, for example, via wireless or wired communication, and is used for inputting the content 10 via a network. The content input processing unit 102 may be an interface for reading content from a recording medium, for example. In this case, the content input processing unit 102 is for inputting content recorded on the recording medium 11.

  The audience rating input unit 103 is an interface for inputting information, and is used for inputting audience rating information 14 of each of a plurality of contents.

  Here, the audience rating information 14 is information indicating how many people who have received the content can view the content. Note that the audience rating information may be input to the preference level calculation device 100 during the broadcast of the content, or may be input to the preference level calculation device 100 after the content broadcast ends. Alternatively, it may be input to the preference calculation device 100 at a predetermined time. This audience rating information 14 is calculated in an external device and is input to the preference level calculation device 100 via the audience rating input unit 103. For example, information provided from an audience rating survey company can be used as audience rating information 14.

  The fashion level calculation unit 104 calculates the fashion level of each of the plurality of contents based on the fashion information 13. The popularity of the content is calculated by using content meta information given to the content or information regarding the content included in the electronic program guide 12.

  Here, the degree of popularity is information indicating how much the content is in line with the world's fashion, or how much the content is suitable for the current season.

  Based on the preference level information 15, the preliminary preference level calculation unit 105 determines the user's preference level for the first content that is the viewing content and the second content that is distributed when the first content is being viewed. Is calculated. The content preference level is calculated by using content meta information given to the content or information regarding the content included in the electronic program guide 12.

  Here, the user's preference level for content is information indicating how much the content matches the user's preference. For example, the user's preference level for the content can be calculated by comparing the user's favorite performer, favorite genre name, and favorite keyword with the content of the content. This preference level information can be obtained as the preference level information 15.

  The character data recognition unit 106 recognizes character data input from the external device 16 and converts it into text data. For example, as shown in FIG. 5, the character data recognition unit 106 recognizes character data such as a telop 301 in the video information of the content displayed on the display device 300 and converts it into text data 302.

  The voice data recognition unit 107 recognizes voice data input from the external device 16 and converts it into text data. For example, as shown in FIG. 5, the audio data recognition unit 107 recognizes audio data such as the announcer's audio 303 in the video information of the content displayed on the display device 300, and converts it into text data 304.

  The keyword extraction unit 108 extracts content meta information attached to content input via the content input processing unit 102, and extracts keywords from the text information of the extracted content meta information using morphological analysis or the like. Alternatively, the keyword extraction unit 108 extracts viewing commentary information on the viewing content, that is, the content viewed by the user from the electronic program guide input via the input unit 101, and morphemes from the extracted program description information text information. Keywords are extracted using analysis or the like. Alternatively, the keyword extracting unit 108 converts the text data in the video information of the content input from the external device 16 into the text data in the text data recognizing unit 106 or the voice data in the text data recognizing unit 107 as text data. And the keyword is extracted from the extracted text information using morphological analysis or the like.

  The keyword preference level calculation unit 109 is configured via the preference level for the content calculated by the preliminary preference level calculation unit 105, the trend level for the content calculated by the trend level calculation unit 104, and the audience rating input unit 103. The preference level for each keyword extracted by the keyword extraction unit 108 is calculated using at least one piece of information regarding the audience rating of the input content.

  The output processing unit 110 associates the preference level of the keyword calculated by the keyword preference level calculation unit 109 with the keyword and outputs the associated keyword to the external device 16.

(Functional configuration of content recommendation device)
FIG. 2 shows a block diagram of a functional configuration of content recommendation apparatus 150 according to the present embodiment. 2, the content recommendation device 150 includes the preference degree calculation device 100 illustrated in FIG. 1, and further includes a reception unit 151, a keyword storage unit 152, a keyword display processing unit 153, a recommended content determination unit 154, and recommended content. An output processing unit 155 is provided.

  The receiving unit 151 is an interface for receiving content broadcasted, for example, wirelessly or via a wire, and is used for inputting the content 10 via a network.

  The keyword storage unit 152 stores the keyword output by the output processing unit 110 in the preference level calculation device 100 and the preference level for the keyword. If the keyword output by the output processing unit 110 already exists in the keyword storage unit 152, the preference level for the keyword output by the output processing unit 110 is updated.

  The keyword display processing unit 153 displays the keyword output by the output processing unit 110 in the preference level calculation device 100 and the preference level for the keyword on the display device 17. The keyword display processing unit 153 displays information stored in the keyword storage unit 152 on the display device 17.

  The recommended content determination unit 154 determines, as recommended content, content that matches the user's preference from among the plurality of contents received by the reception unit 151. The recommended content determination unit 154 determines the recommended content based on the keyword stored in the keyword storage unit 152 and the preference level information about the keyword.

  The recommended content output processing unit 155 outputs the recommended content determined by the recommended content determination unit 154 to the display device 17.

(Functional configuration of content recording / output device)
FIG. 3 shows a block diagram of a functional configuration of the content recording / output device 200 according to the present embodiment. As illustrated in FIG. 3, the content recording / output device 200 includes the preference calculation device 100 illustrated in FIG. 1, and further includes a receiving unit 201, a content recording processing unit 202, a content storage unit 203, and a content output processing unit 204. ing.

  The receiving unit 201 is an interface for receiving content broadcasted, for example, wirelessly or via a wire, and is used for inputting the content 10 via a network.

  The content recording processing unit 202 performs processing for recording content input via the receiving unit 201 and causes the content storage unit 203 to store the content.

  The content output processing unit 204 calls the content stored in the content storage unit 203 and outputs it to the preference calculation device 100.

(Preference level calculation server)
FIG. 4 shows a preference level calculation server 250 according to the present embodiment. As illustrated in FIG. 4, the preference level calculation server 250 includes the preference level calculation device 100 illustrated in FIG. 1, and further includes a viewing history information input unit 251. The preference calculation server 250 is connected to a network, and the viewing history information input unit 251 exchanges information with the external device 16 via the network, and inputs viewing history information of content viewed by the user.

(Configuration of a computer system that implements the preference calculation device 100)
The preference calculation apparatus 100 according to the present embodiment is substantially realized by computer hardware, a program executed by the computer hardware, and data stored in the computer hardware. FIG. 6 shows the internal configuration of the computer system 400.

  Referring to FIG. 6, the computer system 400 includes a computer 420 having an optical disk drive 401 and a magnetic disk drive 402, a monitor 403, and a remote controller (hereinafter referred to as “remote controller”) 404.

  In addition to the optical disk drive 401 and the magnetic disk drive 402, the computer 420 includes a remote control interface (I / F) 405 that receives a signal from the remote control 404, a CPU (central processing unit) 406, a remote control interface 405, a CPU 406, and an optical disk. A bus 407 connected to the drive 401 and the magnetic disk drive 402, a read-only memory (ROM) 408 connected to the bus 407 and storing a boot-up program and the like, and also connected to the bus 407, program instructions, system programs, And a random access memory (RAM) 409 for storing work data and the like.

  Although not shown here, the computer 420 may further include a network adapter board that provides a connection to a local area network (LAN).

  A program for causing the computer system 400 to realize the function of the preference level calculation device 100, the content 10 (see FIG. 1), the electronic program guide 12, the trend information 13, the audience rating information 14, and the preference level information 15 are all optical disks. The data is stored in the optical disk 410 or the magnetic disk 411 inserted into the drive 401 or the magnetic disk drive 402 and further transferred to the hard disk 412. Alternatively, the program, the content 10 (see FIG. 1), the electronic program guide 12, the fashion information 13, the audience rating information 14, and the preference information 15 may be transmitted to the computer 420 through a network (not shown) and stored in the hard disk 412. The program is loaded into the RAM 409 at the time of execution. The program may be loaded directly into the RAM 409 from the optical disk 410, the magnetic disk 411, or via a network.

  This program includes a plurality of instructions that cause the computer 420 to realize the functions of the preference degree calculation apparatus 100 according to the present embodiment. Some of the basic functions necessary to realize these functions are provided by an operating system (OS) or a third-party program running on the computer 420 or various toolkit modules installed on the computer 420. Therefore, this program does not necessarily include all functions necessary for realizing the functions of the preference degree calculation apparatus 100 according to this embodiment. This program includes only instructions that execute control of the above-described preference calculation device 100 by calling an appropriate function or “tool” in a controlled manner so as to obtain a desired result. It only has to be. The operation of computer system 400 is well known and will not be repeated here.

  In addition, the recording medium on which the program for realizing the function of the preference level calculation device 100, the content 10 (see FIG. 1), the electronic program guide 12, the trend information 13, the audience rating information 14, and the preference level information 15 is recorded. , CD-ROM (Compact Disc Read Only Memory), MO (Magneto-Optical Disc), MD (Mini Disc), and DVD (Digital Versatile Disc) optical disc 410, FD (flexible disc) and hard disc 411 such as a hard disc Not limited to. Tapes such as magnetic tapes and cassette tapes, card-type recording media such as IC (Integrated Circuit) cards and optical cards, and semiconductors such as mask ROM, EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), and flash ROM It can be either memory. However, the computer system 400 needs to be equipped with a reading device for reading from these recording media.

(Outline of main operations)
FIG. 7 is a flowchart illustrating an outline of main operations of the preference degree calculation apparatus 100. Referring to FIG. 7, when the preference calculation device 100 is turned on, the program 450 is activated and the process proceeds to step S451.

  In step S451, the keyword extraction unit 108 performs keyword extraction processing, and proceeds to step S452. Details of the keyword extraction process will be described later.

  In step S452, the keyword extraction unit 108 determines whether the keyword extraction processing in step S451 has succeeded. If the keyword has been successfully extracted, the process proceeds to step S453, and if it has failed, the process is terminated.

  In step S453, at least one of the audience rating input unit 103, the fashion level calculation unit 104, and the preliminary preference level calculation unit 105 calculates an evaluation value related to the preference for the content, and the process proceeds to step S454. Details of the content evaluation value calculation / extraction process will be described later.

  In step S454, it is determined whether at least one of the audience rating input unit 103, the fashion level calculation unit 104, and the preliminary preference level calculation unit 105 has succeeded in the content evaluation value calculation / extraction process in step S453. If the content evaluation value calculation / extraction process is successful, the process proceeds to step S456. If the content evaluation value calculation / extraction process is unsuccessful, the process proceeds to step S455.

  In step S455, at least one of the audience rating input unit 103, the fashion level calculation unit 104, and the preliminary preference level calculation unit 105 sets an evaluation value related to the preference for content to a default value, and the process proceeds to step S456. For example, the trend level calculation unit 104 sets a default trend level as an evaluation value related to the preference for content. In addition, the evaluation value regarding the preference for the default content is not limited. For example, “there is no evaluation value regarding the preference for content” may be used. Furthermore, the evaluation value related to the preference for the default content may be held in advance by at least one of the audience rating input unit 103, the fashion level calculation unit 104, and the preliminary preference level calculation unit 105, or the user himself You may input default information.

  In step S456, the keyword preference level calculation unit 109 calculates the preference level for the keyword extracted in step S451, and the process ends. Details of the preference level calculation process will be described later.

(Keyword extraction process)
FIG. 8 is a flowchart showing the keyword extraction process executed in step S451 shown in FIG. Referring to FIG. 8, when keyword extraction processing 500 is started, in step S501, it is determined whether or not keyword extraction unit 108 extracts keywords from content meta information input via content input processing unit 102. To do. This determination may be performed based on an instruction from the user, or may be performed by extracting a keyword when content meta information is present. If a keyword is extracted from the content meta information, the process proceeds to step S502. Otherwise, the process proceeds to step S503. If the content meta information is not input via the content input processing unit 102, the process proceeds to step S503.

  In step S502, the keyword extraction unit 108 extracts keywords from the content meta information input via the content input processing unit 102, and the process proceeds to step S503.

  In step S <b> 503, the keyword extraction unit 108 determines whether to extract keywords from the electronic program guide input via the input unit 101. This determination may be performed based on an instruction from the user, or when an electronic program guide exists, a keyword may be extracted. If a keyword is extracted from the electronic program guide, the process proceeds to step S504. Otherwise, the process proceeds to step S505. If the electronic program guide is not input via the input unit 101, the process proceeds to step S505.

  In step S504, the keyword extracting unit 108 extracts a keyword from the electronic program guide including information related to the content input via the input unit 101, and the process proceeds to step S505.

  In step S505, the character data recognition unit 106 determines whether to extract a keyword from the character data in the content video information. This determination may be performed based on an instruction from the user, or a keyword may be extracted when the character data recognition unit 106 can recognize character data. If a keyword is extracted from the character data in the content video information, the process proceeds to step S506. Otherwise, the process proceeds to step S508.

  In step S506, the character data recognition unit 106 recognizes the character data from the video information of the content and converts it into text data, and the process proceeds to step S507.

  In step S507, the keyword extraction unit 108 extracts keywords from the text data converted in step S506, and the process proceeds to step S508.

  In step S508, the audio data recognition unit 107 determines whether to extract a keyword from the audio data included in the content. This determination may be performed based on an instruction from the user, or a keyword may be extracted when the speech data recognition unit 107 recognizes speech data. If a keyword is extracted from the audio data, the process proceeds to step S509. If not, the process proceeds to step S511.

  In step S509, the voice data recognition unit 107 recognizes voice data included in the content and converts it into text data, and the process proceeds to step S510.

  In step S510, the keyword extraction unit 108 extracts keywords from the text data converted in step S509, and the process proceeds to step S511.

  In step S511, the keyword extraction unit 108 determines whether to extract a keyword from other information. This determination is made based on whether or not other information related to the content has been input. When other information is input and a keyword is extracted from the information, the process proceeds to step S512. When not extracted, the process ends.

  Here, the other information includes, for example, information such as a keyword input by the user himself / herself while viewing the viewing content, but is not necessarily limited thereto.

  In step S512, the keyword extraction unit 108 extracts keywords from other information, and the process ends.

(Content evaluation value calculation / extraction process)
FIG. 9 is a flowchart showing the content evaluation value calculation / extraction process executed in step S453 shown in FIG. Referring to FIG. 9, when content evaluation value calculation / extraction process 550 is started, preliminary preference calculation unit 105 in step S551 broadcasts the user while viewing the viewing content, and calculates the preference. It is determined whether or not user preference information for content that can be received by the device including the apparatus 100 is calculated. This determination may be performed based on an instruction from the user, or when the preference level information 15 is input, the preference level information may be calculated. If the user preference information is to be calculated, the process proceeds to step S552, and if not, the process proceeds to step S553.

  In step S552, the preliminary preference level calculation unit 105 includes at least the input preference level information 15, the electronic program guide 12 input via the input unit 101, and the content meta information input via the content input processing unit 102. Based on one, the preference level of the content is calculated as an evaluation value related to the preference for the content.

  In step S553, the trend level calculation unit 104 determines whether to calculate the trend level of content that is broadcast while the user is viewing the viewing content and can be received by a device including the preference level calculation device 100. To do. This determination may be performed based on an instruction from the user, or when the fashion information 13 is input, the fashion level may be calculated. If the fashion level is to be calculated, the process proceeds to step S554, and if not, the process proceeds to step S555.

  In step S554, the trend level calculation unit 104 outputs at least one of the input trend information 13, the electronic program guide 12 input via the input unit 101, and the content meta information input via the content input processing unit 102. Based on the above, the popularity level of content is calculated as an evaluation value related to the preference for content.

  In step S555, the audience rating input unit 103 determines whether or not to extract the audience rating of content that is broadcast while the user is viewing the viewing content and can be received by a device including the preference calculation device 100. To do. This determination may be performed based on an instruction from the user, or when the audience rating information 14 is input, the audience rating may be extracted. If the audience rating is extracted, the process proceeds to step S556, and if not extracted, the process ends.

  In step S556, the audience rating input unit 103 extracts the audience rating of the content as an evaluation value related to the preference for the content from the audience rating information 14, and ends the process.

(Preference level calculation processing)
FIG. 10 is a flowchart showing the preference level calculation process executed in step S456 shown in FIG. Referring to FIG. 10, when the preference level calculation process 600 is started, in step S601, the keyword preference level calculation unit 109 performs steps S502, S504, S507, S510, and S512 shown in FIG. At least one of the keywords extracted in step S552, step S554, and step S556 shown in FIG. 9 is used, or the evaluation value related to the preference for the extracted content is used. Is calculated, and the process proceeds to step S602.

  In step S602, the keyword preference level calculation unit 109 determines whether or not the preference level calculation in step S601 has succeeded. If the preference degree has been successfully calculated, the process proceeds to step S604, and if it has failed, the process proceeds to step S603.

  In step S603, the keyword preference calculation unit 109 sets the preference for the keyword to a value determined by default, and the process proceeds to step S604. The default preference level may be stored in advance by the keyword preference level calculation unit 109, or may be input by the user himself / herself.

  In step S604, the output processing unit 110 outputs the keyword for which the keyword preference level calculation unit 109 has calculated the preference level and the preference level for the keyword to the external device 16, and the process proceeds to step S605.

  In step S605, the keyword preference calculation unit 109 calculates the preference for all the keywords extracted in at least one of step S502, step S504, step S507, step S510, and step S512 shown in FIG. Determine whether or not. If the preference level has not been calculated for all keywords, the process returns to step S601, and if calculated, the process ends.

  Hereinafter, examples 1 to 5 will be exemplified as specific operations of the preference degree calculation apparatus 100.

(Example 1)
In Example 1 shown below, when a user is viewing content, the user's preference information for content that is being broadcast at the same time, including the viewing content, is used as a keyword extracted from the viewing content. The operation of the preference level calculation device 100 when calculating the preference level is illustrated.

  FIG. 11 shows a content group 620 broadcast from “first channel” to “third channel” from 19:00 to 21:00. In FIG. 11, along with the flow of the time zone 621, the names of contents broadcast on each of “first channel” 622, “second channel” 623, and “third channel” 624 are shown. For example, in “first channel”, “content A” is broadcast from 19:00 to 20:00, and “content D” is broadcast from 20:00 to 21:00. However, it is assumed that there are currently three channels that can be received by the device including the preference calculation device 100, from “first channel” to “third channel”.

  FIG. 12A to FIG. 12C show electronic program guides for the programs “content A” to “content C”. 12A shows an electronic program guide 640 for “content A”, FIG. 12B shows an electronic program guide 641 for “content B”, and FIG. 12C shows an “electronic program guide for content C” 642. The electronic program guide for each content includes information on the program ID, broadcast start date, start time, broadcast end date, end time, channel, genre, program name, and program content. For example, in the electronic program guide 640 of “content A”, the program ID is “S200401419000531”, and “10:00” of “October 14, 2004” to “20:00 of“ October 14, 2004 ”. ”Shows that the program“ News 19:00 ”classified into the genre“ Society / Report ”will be broadcast on“ First Channel ”. However, information included in the electronic program guide of each content is not necessarily limited to these. Furthermore, the contents shown in FIGS. 12A to 12C always include the above information. However, for example, some contents may not include program contents. Note that these electronic program guides are distributed from the Internet, broadcasting stations, and the like, and are input to the preference calculation device 100 via the input unit 101.

  Here, in the case where user X views “content C” of “third channel” from 19:00 in content group 620 being broadcast as shown in FIG. 11, according to flowchart 500 shown in FIG. The keyword extraction process from “content C” will be described.

  In step S501, the keyword extraction unit 108 does not extract the keyword from the content meta information because the content meta information is not input via the content input processing unit 102, and the process proceeds to step S503.

  In step S503 and step S504, the keyword extraction unit 108 extracts keywords from the electronic program guide 642 of “content C” shown in FIG.

  However, the method for extracting keywords from the electronic program guide is not limited. For example, morphological analysis of the program name and program content may be performed, and only the noun part may be extracted as a keyword therefrom. Further, the keyword to be extracted may be one or plural. Furthermore, after extracting a plurality of keywords and presenting them to the user, the keyword to be extracted may be determined according to a response from the user.

  Here, it is assumed that two keywords “hot spring” and “travel” are extracted.

  In step S505 to step S512, the keyword extraction unit 108 does not extract the keyword from each information because the corresponding information does not exist, and ends the keyword extraction process.

  Note that the keyword extraction method is not limited to extraction from the electronic program guide as described above. For example, the keywords may be extracted from the content meta information given to the viewing content using the content meta information related to the content given to the content shown in FIGS. 13 (a) to 13 (c).

  13A, the content meta information 660 assigned to “content A”, the content meta information 661 assigned to “content B” in FIG. 13B, and “content C” in FIG. 13C. The assigned content meta information 662 is shown, and each piece of content meta information includes information on broadcast date and time, channel, performer, genre, and keyword.

  For example, in the content meta information given to “content A”, content belonging to the genre of “society / report” from “19:00 to 20:00” on “October 14, 2004” is “first channel”. The content's performer is “Tetsuya Kurume” and the keywords of the content are “Major League”, “Juro”, “Politics”, “Election of the House of Representatives”, and “Tetsuya Kurume” Yes. However, the information included in the content meta information given to each content is not necessarily limited to these.

  The content meta information is distributed from the Internet or a broadcasting station together with the content, and is input to the preference level calculating device 100 through the content input processing unit 102 together with the content. In addition, when content recorded on the recording medium 11 is input to the preference calculation device 100 via the content input processing unit 102, content meta information is recorded on the recording medium 11 together with the content, and together with the content. Entered.

  Here, since the user X is viewing “content C” of “third channel” from 19:00 in the content group 620 being broadcast as shown in FIG. 11, the method described above is used. Keywords are extracted according to the flowchart shown in FIG. In step S501 and step S502, the keyword extraction unit 108 extracts a keyword from the content meta information 662 assigned to “content C”. For example, the keyword extraction unit 108 extracts keywords “Burari Journey”, “Secret Hot Spring”, “Hot Spring”, “Ryokan”, and “Taste”.

  Furthermore, the keyword extraction method is not limited to the method described above. For example, as shown in FIG. 14, character data in video information of viewing content may be recognized and converted into text data, and keywords may be extracted from the text data.

  FIG. 14 shows a case where “content C” is displayed on display device 680 and a keyword is extracted from character data 681 included in video information of “content C”. In this case, the character data recognition unit 106 recognizes the character data 681 and converts it into text data 682.

  Here, since the user X is viewing “content C” of “third channel” from 19:00 in the content group 620 being broadcast as shown in FIG. 11, the method described above is used. Keywords are extracted according to the flowchart shown in FIG. In step S505, step S506, and step S507, the character data recognition unit 106 recognizes the character data 681 included in the video information of “content C” and converts it into text data 682, and the keyword extraction unit 108 reads the text. A keyword is extracted from the data 682. For example, the keyword extraction unit 108 extracts the keywords “secret hot spring” and “◯ △ hot spring”.

  Furthermore, the keyword extraction method is not limited to the method described above. For example, as shown in FIG. 14, audio data included in viewing content may be recognized and converted into text data, and keywords may be extracted from the text data.

  FIG. 14 shows a case where “content C” is displayed on display device 680 and a keyword is extracted from audio data 683 included in “content C”. In this case, the voice data recognition unit 107 recognizes the voice data 683 and converts it into text data 684.

  Here, since the user X is viewing “content C” of “third channel” from 19:00 in the content group 620 being broadcast as shown in FIG. 11, the method described above is used. Keywords are extracted according to the flowchart shown in FIG. In step S508, step S509, and step S510, the voice data recognition unit 107 recognizes the voice data 683 included in the “content C” and converts it into text data 684, and the keyword extraction unit 108 uses the keyword from the text data 684 as a keyword. To extract. For example, the keyword extracting unit 108 extracts the keyword “◯ △ hot spring”.

  Next, FIG. 9 shows content evaluation value calculation / extraction processing in the case where the user X is viewing or listening to “content C” at 19:00 to calculate or extract an evaluation value related to the preference for content. A description will be given according to the flowchart 550.

  Here, it is assumed that the preference level of “contents A and B” being broadcast simultaneously with “content C” and the preference level of “content C” as viewing content are extracted as evaluation values regarding the preference for content.

  In step S551 and step S552, the preference level of “content C” as the viewing content and the preference levels of “content A” and “content B” broadcast simultaneously are calculated.

  Here, FIG. 15 shows user X's preference information 700. FIG. 15 shows user X's favorite genre 701 and user X's favorite keyword 702. When the preference level of each content is calculated using the content meta information given to each content shown in FIGS. 13A to 13C, the preliminary preference level calculation unit 105 shown in FIG. From the content meta information input from the processing unit 102 and the preference information 700 of the user X, the preference level of “contents A to C” is calculated. Here, when the word registered in the favorite genre 701 in the preference information 700 of the user X is included in the content meta information, the point registered in the favorite keyword 702 in the favorite information 702 in the user X preference information 700 is 10 points. If it is included in the content meta information, it is 20 points.

  In the case of “content A”, “society / report” matches in genre and “major league” and “juro” match in keyword, so the preliminary preference calculation unit 105 sets the preference of “content A” to 50 calculate. Similarly, the preliminary preference level calculation unit 105 calculates the preference level of “content B” as 80 and the preference level of “content C” as 30. Note that the preference degree calculation method and expression method are not necessarily limited thereto.

  In step S553, since the trend level calculation unit 104 does not calculate the trend level of the content, the process proceeds to step S555. In step S555, the audience rating input unit 103 does not extract the content rating rate, so the content evaluation value calculation / extraction process is performed. Exit.

  Next, using the preference information of “contents A” and “contents C” being broadcast simultaneously with “contents C” and “contents C”, it is extracted from the electronic program table 642 related to “contents C” shown in FIG. The preference level calculation processing for calculating the preference levels for the two keywords “hot spring” and “travel” will be described with reference to the flowchart 600 shown in FIG.

  In step S601, the keyword preference level calculation unit 109 calculates the preference level for each keyword extracted in step S504 using the preference level information of the content being broadcast (contents A to C) calculated in step S552. To do. Here, as a method for calculating the preference level, a calculation formula 720 which is the preference level calculation formula A shown in FIG. 16 is used.

  The preference level calculation formula A is 0.1 for the difference between the user's preference level for the viewing content from which the keyword is extracted and the preference level of the content having the highest preference level among the content broadcast while viewing the viewing content. A value obtained by adding 10 to the value multiplied by is set as the preference degree for the keyword. That is, the intent of the preference level calculation formula A is that when content with a higher preference level than the viewing content is broadcast on the back, the preference level of the viewing content and the preference level of the content being broadcast on the back The greater the difference from the highest preference level, the higher the preference level can be set for keywords extracted from the viewing content.

  However, the preference level calculation method is not necessarily limited to this. For example, a table for converting the degree of preference for content into the degree of preference for the keyword may be prepared.

  Here, the preference level of “content C” from which two keywords “hot spring” and “travel” are extracted is 30, and the preference level is the highest among the contents broadcast while viewing “content C”. The degree of preference for high content is 80 for “content B”. Accordingly, the keyword preference level calculation unit 109 calculates the preference level by substituting these values into the preference level calculation formula A. In this case, the preference level is calculated as 15.

  In step S <b> 604, the output processing unit 110 outputs information that the preference level of the keywords “hot spring” and “travel” is 15 to the external device 16.

  In step S605, the keyword preference level calculation unit 109 determines whether there is a keyword for which the preference level has not been calculated. If there is a keyword for which the preference level has not been calculated, the process returns to step S601, When the preference level is calculated for the keyword, the preference level calculation process is terminated.

  Here, a case where the external device 16 includes a keyword output unit 110 and a keyword storage unit for storing the degree of preference for the keyword will be described with reference to FIG.

  FIG. 17 shows a state before and after updating the keyword storage unit displayed in association with the keyword and the degree of preference for the keyword. Since the keyword “hot spring” already exists in the keyword storage unit, the preference level for the updated keyword “hot spring” is updated to 20 by adding the preference level 15. On the other hand, since the keyword “travel” does not exist in the keyword storage unit, the degree of preference for the keyword “travel” becomes 15 by adding a new keyword.

  Furthermore, in the content recommendation device 150 including the preference calculation device 100, the keyword storage unit 152, and the recommended content determination unit 154 shown in FIG. 2, the keywords stored in the keyword storage unit shown in FIG. By using the degree of preference, it is possible to determine the content recommended to the user. Since the preference level for the keyword stored in the keyword storage unit is calculated according to the content that was broadcast on the back side when the user is viewing the content, a value suitable for the user's preference is recorded. ing.

  Therefore, the recommended content determination unit 154 selects the content having the highest preference point as the user's preference point for the content, for example, when the total preference of each keyword extracted from the content that is a candidate for recommendation to the user is used. Determined as recommended content for As a result, the recommended content determination unit 154 can recommend content that matches the user's preference.

  Further, FIG. 18 shows a state in which the keyword extracted from “content C” and the preference level of the keyword are displayed on the display device. As a result, the user can grasp the degree of preference for the keyword extracted from the viewing content.

  In the calculation of the preference level of the keyword using the preference level calculation formula A, the preference level can be increased as the keyword is extracted from the content with the lower preference level. This is because the keyword extracted from the viewing content is more representative of the user's preference as the content with lower preference is viewed.

  As described above, in the configuration of the first embodiment, the keyword extracted from the viewing content using the user's preference information for the content being broadcast when the user is viewing the viewing content. The degree of preference can be calculated.

(Example 2)
In Example 2 shown below, when a user views content, the trend level information about the content being broadcast at the same time including the viewing content is used to extract keywords from the viewing content. The operation of the preference level calculation device 100 when calculating the preference level is illustrated.

  Here, when the user X views “content A” of “first channel” from 19:00 in the broadcast content group 620 as shown in FIG. 11, it is extracted from “content A”. A preference level calculation process for calculating a preference level for the keyword using the trend information will be described.

  Note that the keyword extraction method for “content A”, which is the viewing content, may be any method shown in the above-described embodiment. Here, it is assumed that the keywords are extracted from the keywords included in the content meta information 660 of “content A” shown in FIG. 13A, and the keyword extraction unit 108 is “major league”, “juro”, “politics”, “the House of Representatives” It is assumed that five keywords “election” and “Tetsuya Kurume” are extracted.

  First, the flowchart shown in FIG. 9 shows the content evaluation value calculation / extraction process in the case where the evaluation value related to the preference for content is calculated or extracted while the user X is viewing “content A” at 19:00. Description will be made according to 550.

  Here, it is assumed that the popularity of “contents B and C” being broadcast simultaneously with “content A” and the popularity of “content A” as viewing content are extracted as evaluation values related to the preference for content.

  In steps S551 and S552, the preliminary preference level calculation unit 105 does not calculate the content preference level, and thus the process proceeds to step S553.

  In step S553 and step S554, the trend level calculation unit 104 calculates the trend level of “content A” as the viewing content, and the trend levels of “content B” and “content C” broadcast simultaneously.

  Here, the trend information 740 is shown in FIG. In FIG. 19, a keyword 741 and a trend point 742 obtained by digitizing the trend of the keyword are displayed in association with each other. The trend point is assumed to be in line with the current trend as the numerical value increases, and this information is created in an external device and input to the preference degree calculation device 100 as the trend information 13.

  The trend points are created, for example, as follows. FIG. 22 shows an example of a functional configuration block diagram of a trend point calculation device 800 that calculates trend points.

  A trend point calculation device 800 illustrated in FIG. 22 includes a reception unit 801, an input unit 802, a keyword extraction processing unit 803, a keyword accumulation management unit 804, a trend point conversion unit 805, and a trend point output processing unit 806.

  The content includes not only video programs recorded on television programs and recording media broadcasted via wireless or wired communication, but also music media such as radio programs and music provided in the same manner. That is, it refers to all multimedia contents provided by various forms and various media.

  The receiving unit 801 is an interface that receives content broadcasted, for example, via wireless or wired communication, and is used for inputting the content 810 via a network.

  The input unit 802 is an interface for reading content from a recording medium, for example, and is used for inputting content recorded on the recording medium 811.

  The keyword extraction processing unit 803 extracts keywords included in content input via the receiving unit 801 or the input unit 802. For example, keywords are extracted from content meta information given to the content.

  The keyword accumulation management unit 804 manages the keywords extracted by the keyword extraction processing unit 803. For example, the keyword accumulation management unit 804 manages a value obtained by counting the frequency of occurrence of each keyword and the keyword in association with each other, and further manages a value associated with the accumulated broadcast time of content related to the keyword. It is a configuration.

  The trend point conversion unit 805 converts the information managed by the keyword accumulation management unit 804, for example, a value obtained by counting the appearance frequency of each keyword, or a value obtained by accumulating the broadcast time for each keyword into a trend point for each keyword. To do.

  The trend point output processing unit 806 outputs the trend point of each keyword converted by the trend point conversion unit 805 to the external device 812.

  Thus, for example, in broadcasting for a predetermined period from one or a plurality of broadcasting stations, keywords included in the broadcasted information are extracted, and the frequency of appearance of the keywords is counted and converted as trend points. Alternatively, the trend points can be calculated by accumulating the broadcast time for the keyword and converting it as a trend point.

  Further, the above-described configuration for obtaining the fashion points (fashion information) may be included in the fashion calculation unit 104. In this case, instead of the trend information 13, reception information from the broadcast music is input to the trend level calculation unit 104.

  Here, the popularity of each content is the sum of trend points corresponding to all keywords included in the content meta information given to each content shown in FIGS.

  In the case of “Content A”, the trend point of “Meeting of the House of Representatives” is 5, the trend point of “Major League” and “Juro” is 3, and the trend point of “Politics” and “Tetsuya Kurume” is 1. The degree calculation unit 104 calculates the popularity degree of “content A” as 13. Similarly, the trend level calculation unit 104 calculates the trend level of “content B” as 27 and the trend level of “content C” as 10. Note that the fashion level calculation method and expression method are not necessarily limited thereto.

  In step S555, since the audience rating input unit 103 does not extract the audience rating of the content, the content evaluation value calculation / extraction process ends.

  Next, using the popularity information of “contents B” and “contents A” simultaneously broadcast with “contents A” and “contents A”, “major league”, “juro”, “ The preference level calculation processing for calculating the preference levels for the five keywords “politics”, “election of the House of Representatives”, and “Tetsuya Kurume” will be described with reference to the flowchart 600 shown in FIG.

  In step S601, the keyword preference level calculation unit 109 calculates the preference level for each keyword extracted in step S502, using the popularity level of the content being broadcast (contents A to C) calculated in step S554. . Here, as a method for calculating the preference level, a calculation formula 721 that is a preference level calculation formula B shown in FIG. 16 is used.

  The preference level calculation formula B sets 0.1 as the difference between the popularity level of the viewing content from which the keyword is extracted and the popularity level of the content that is broadcasted while viewing the viewing content. A value obtained by adding 10 to the multiplied value is set as a preference degree for the keyword. That is, the intent of the preference level calculation formula B is that when content with a higher fashion level than the viewing content is broadcast on the back side, the popularity level of the viewing content and the fashion level of the content being broadcast on the back side The greater the difference from the highest fashion level, the higher the preference level for keywords extracted from the viewing content.

  However, the preference level calculation method is not necessarily limited to this. For example, a table for converting the trend level into the preference level for the keyword may be prepared.

  Here, the popularity of “Content A” from which the five keywords “major league”, “juro”, “politics”, “election of the House of Representatives”, and “Tetsuya Kurume” are extracted is “content A”. The most popular content of the content broadcast during viewing is “content B” with 27. Accordingly, the keyword preference level calculation unit 109 calculates the preference level by substituting these values into the preference level calculation formula B. In this case, the preference level is calculated as 11.4.

  In step S604, the output processing unit 110 receives information indicating that the preference degree of the five keywords “major league”, “juro”, “politics”, “election of the House of Representatives”, and “Tetsuya Kurume” is 11.4. Output to the device 16.

  In step S605, the keyword preference level calculation unit 109 determines whether there is a keyword for which the preference level has not been calculated. If there is a keyword for which the preference level has not been calculated, the process returns to step S601, When the preference level is calculated for the keyword, the preference level calculation process is terminated.

  Note that the degree of preference of each keyword may be calculated in consideration of the trend points of each keyword shown in FIG. For example, if you use a preference level calculation formula that suppresses the preference level for highly popular keywords, you can reduce the preference level for trendy keywords even for “watching” that you are watching recently. Can respond.

  In the calculation of the keyword preference level using the preference level calculation formula B, it is possible to increase the preference level as the keyword is extracted from the content having a lower trend. This is based on the idea that as the content with a lower fashion level is viewed, the content is viewed without being influenced by the fashion, and the keywords extracted from the viewed content well represent the user's preference.

  As described above, in the configuration of the second embodiment, the degree of preference for the keyword extracted from the viewing content using the trend information on the content being broadcast when the user is viewing the viewing content. Can be calculated.

(Example 3)
In Example 3 shown below, when a user is viewing content, the viewing rate information of the content being broadcast at the same time including the viewing content is used for keywords extracted from the viewing content. The operation of the preference level calculation device 100 when calculating the preference level is illustrated.

  Here, when the user X views “content C” of “third channel” from 19:00 in the content group 620 being broadcast as shown in FIG. 11, it is extracted from “content C”. The preference for the keyword using the audience information of “content A”, “content B”, and “content C”, which are contents that are currently broadcast and can be received by the device including the preference calculation device 100 The preference level calculation process when calculating the level will be described with reference to the flowchart 600 shown in FIG.

  It should be noted that the keyword extraction method for “content C”, which is the viewing content, may be any method shown in the above-described embodiment. Here, it is assumed that the keywords are extracted from the keywords included in the content meta information 662 of “content C” shown in FIG. 13C, and the keyword extraction unit 108 performs “burari trip”, “hidden hot spring”, “hot spring”, “ It is assumed that five keywords “inn” and “taste” are extracted.

  Here, the audience rating of the content being broadcast is input as the audience rating information 14 to the preference degree calculation apparatus 100 via the audience rating input unit 103, and the audience rating of “content A” is 8%, “content B” ”Is 16%, and“ Content C ”is 12%. Note that the audience rating of content being broadcast is calculated by an external device.

  In step S601, the keyword preference level calculation unit 109 calculates the preference level for each keyword extracted in step S502, using the audience rating extracted in step S556. Here, as a method for calculating the preference level, a calculation formula 722 which is a preference level calculation formula C shown in FIG. 16 is used.

  The preference level calculation formula C is a value obtained by adding 10 to a value obtained by multiplying the difference between the viewing rate of the viewing content from which the keyword is extracted and the highest viewing rate among the currently receivable content by adding 10 to the preference for the keyword. Degree. That is, the intent of the preference level calculation formula C is that, when content with a higher audience rating than the viewing content is broadcast on the back, the viewing rate of the viewing content and the viewing rate of the content being broadcast on the back The greater the difference from the highest audience rating, the higher the preference level can be set for keywords extracted from the viewing content.

  However, the preference level calculation method is not necessarily limited to this. For example, a table for converting the audience rating into the degree of preference for the keyword may be prepared.

  Here, the audience rating of “Content C” from which five keywords “Burari Journey”, “Secret Hot Spring”, “Hot Spring”, “Ryokan”, and “Taste” are extracted is 12% and is currently being broadcast. The viewing rate of the content having the highest viewing rate among existing and receivable content is 16% of “content B”. Accordingly, the keyword preference level calculation unit 109 calculates the preference level by substituting these values into the preference level calculation formula C. In this case, the preference level is calculated as 18.

  In step S <b> 604, the output processing unit 110 informs the external device 16 that the preference level of the five keywords “Buri Journey”, “Secret Hot Spring”, “Hot Spring”, “Ryokan”, and “Taste” is 18. Output.

  In step S605, the keyword preference level calculation unit 109 determines whether there is a keyword for which the preference level has not been calculated. If there is a keyword for which the preference level has not been calculated, the process returns to step S601, When the preference level is calculated for the keyword, the preference level calculation process is terminated.

  In the calculation of the keyword preference level using the preference level calculation formula C, the preference level can be increased as the keyword is extracted from the content having a lower audience rating. This is because the keyword extracted from the viewing content better represents the user's preference as the content with a lower audience rating is viewed.

  As described above, in the configuration of the third embodiment, the degree of preference for the keyword extracted from the viewing content using the viewing rate information of the content broadcast when the user is viewing the viewing content. Can be calculated.

(Example 4)
In Example 4 shown below, when a user views a recorded content without viewing the content that is currently being broadcast and can be received, the recorded content that is the viewing content is used by using an evaluation value related to the preference for the content. The operation of the preference level calculation device 100 when calculating the preference level for the keywords extracted from FIG. Here, as shown in FIG. 3, there is a content recording / output device 200 including a preference calculation device 100 and a content storage unit 203, and “content F” is stored in the content storage unit 203 as shown in FIG. It is assumed that

  FIG. 20 shows content meta information 760 assigned to “content F”, and the content meta information includes information on each of broadcast date / time, channel, performer, genre, and keyword. For example, in the content meta information assigned to “content F”, content belonging to the “drama” genre is broadcast on “third channel” from “21:00 to 22:00” on “October 8, 2004”. The performer of the content is “Tetsuro Takeda”, and the keywords of the content are “Gakuen”, “Education Problem”, and “Kinkyu-sensei”. However, the information included in the content meta information given to each content is not necessarily limited to these. The content meta information is distributed together with the content from the Internet or a broadcasting station, and is stored and managed together with “content F” when “content F” is stored in the content storage unit 203. Also, when content recorded on a recording medium or the like is input via the receiving unit 201, content meta information is recorded on the recording medium or the like together with the content and is managed in the content storage unit 203 together with the content.

  Here, in the content group 620 broadcasted as shown in FIG. 11, the user X has recorded in the past shown in FIG. 20 without viewing any content broadcasted in any channel from 19:00. A flowchart 600 shown in FIG. 10 shows a preference level calculation process when calculating a preference level for a keyword extracted from “content F” when “content F” stored in the content storage unit 203 is viewed. It explains according to.

  Note that the method for extracting the keyword related to “content F” as the viewing content may be any method shown in the above-described embodiment. Here, it is assumed that keywords are extracted from the keywords included in the content meta information 760 of “content F” shown in FIG. 20, and the keyword extraction unit 108 has three types of “school”, “education problem”, and “Kinkyu teacher”. Assume that keywords are extracted.

  Here, it is assumed that the viewing rate of the currently broadcast content is used as an evaluation value regarding the preference for the content while the recorded content is viewed, the viewing rate of “content A” is 8%, and the viewing rate of “content B” is It is assumed that the viewing rate of 16% and “content C” is 12%. Furthermore, it is assumed that the viewing rate of “content F” when “content F” was broadcast in the past was 14%. Note that the audience ratings of “content A”, “content B”, “content C”, and “content F” are calculated by an external device and input as audience rating information 14 via the audience rating input unit 103.

  In step S601, the keyword preference level calculation unit 109 calculates the preference level for each keyword extracted in step S502, using the audience rating extracted in step S556. Here, as a method for calculating the preference level, a calculation formula 722 which is a preference level calculation formula C shown in FIG. 16 is used.

  However, the preference level calculation method is not necessarily limited to this. For example, a table for converting the audience rating information into the degree of preference for the keyword may be prepared.

  Here, the viewing rate of “Content F” from which three keywords “Gakuen”, “Education Problem”, and “Kinkyu-sensei” are extracted is 14%, and the content that is currently being broadcast and can be received is Among them, the viewing rate of the content with the highest viewing rate is 16% of “Content B”. Accordingly, the keyword preference level calculation unit 109 calculates the preference level by substituting these values into the preference level calculation formula C. In this case, the preference level is calculated as 14.

  In step S604, the output processing unit 110 outputs to the external device 16 information indicating that the preference levels of the three keywords “school”, “education problem”, and “Kinku-sensei” are 14.

  In step S605, the keyword preference level calculation unit 109 determines whether there is a keyword for which the preference level has not been calculated. If there is a keyword for which the preference level has not been calculated, the process returns to step S601, When the preference level is calculated for the keyword, the preference level calculation process is terminated.

  In the fourth embodiment, the case where “content F”, which has been recorded in the past, is viewed has been described. However, the present invention is not necessarily limited to this. For example, the content recorded on a recording medium such as a DVD rented from a rental store may be viewed. In this case, the content meta information related to the content recorded on the recording medium is recorded on the recording medium together with the content. In addition, when viewing content recorded on a recording medium such as a DVD rented from a rental store or the like, the audience rating cannot be used as an evaluation value regarding the preference for the content. Etc. is used to calculate the degree of preference for the keyword extracted from the content.

  In the fourth embodiment, the audience rating information of the content broadcasted while viewing “content F” is used as the evaluation value regarding the preference for the content. However, the rating value is not necessarily limited thereto. The user's preference level or the popularity level of content may be used, or a combination of these may be used.

  As described above, in the configuration of the fourth embodiment, even when the user views the recorded content or the like without viewing the currently broadcasted and receivable content, the evaluation value regarding the preference for the content is obtained. It is possible to calculate a preference level for a keyword extracted from viewing content such as recorded content.

(Example 5)
In Example 5 described below, a case where a server provided with a preference degree calculation device is used will be described.

  FIG. 21 shows a state in which a display device 780, a server 781 having a preference calculation device, and a keyword management device 782 are connected via a network as client machines used by a user.

  The network may be wired or wireless.

  In this case, when the user views the content on the display device 780, the viewing information is transmitted to the server 781 through the network. Based on the viewing information transmitted from the display device 780, the server 781 calculates the degree of preference for the keyword extracted from the viewing content of the user. This processing is performed by the method described in the above embodiment.

  Information on the preference level for the keyword calculated by the server 781 is sent to the keyword management device 782, and the keyword management device 782 manages the keyword and the preference level for the keyword.

  Further, the server 781 may be provided with a recommended content determination function for determining recommended content for the user. In this case, the server 781 determines recommended content for the user based on the keyword managed by the keyword management device 782 and the degree of preference for the keyword, and presents it on the display device 780.

  As described above, in the configuration of the fifth embodiment, by using a server provided with a preference level calculation device, all processing can be performed on the server side, and the processing burden on the user terminal can be reduced.

(Modified Example)
In the embodiment described above, keywords are extracted in units of content, such as electronic program guides in units of content and content meta information assigned to the units of content, but the content is composed of one or more parts depending on the content. If so, keywords may be extracted in part units.

  In this case, the preference level for the keyword is extracted using at least one of the preference level information of the part unit, the popularity level information of the part unit, and the audience rating information of the part unit for the keyword extracted by the part unit. May be calculated.

  That is, if one part in the above-described embodiment is regarded as one content, the same effect can be obtained.

  The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

  The present invention can also be applied to uses such as a preference level calculation device that extracts a keyword from viewing content and obtains the user's preference for the keyword, or an information recommendation device that recommends information that matches the preference to the user. it can.

It is a block diagram which shows the functional structure of the preference calculation apparatus which concerns on one embodiment of this invention. It is a block diagram which shows the functional structure of the content recommendation apparatus provided with the preference calculation apparatus which concerns on one embodiment of this invention. It is a block diagram which shows the functional structure of the content recording output device provided with the preference calculation apparatus which concerns on one embodiment of this invention. It is a block diagram which shows the functional structure of the preference calculation server provided with the preference calculation apparatus which concerns on one embodiment of this invention. It is explanatory drawing which shows the recognition operation | movement of the character data and audio | voice data which are contained in the content currently displayed on the display apparatus in the character data recognition part and audio | voice data recognition part which were shown in FIG. It is a block diagram which shows the structure of the computer system which implement | achieves each function part of the preference degree calculation apparatus which concerns on this Embodiment. It is a flowchart which shows the main operation | movement of the preference degree calculation apparatus shown in FIG. It is a flowchart which shows the keyword extraction process performed by step S451 of FIG. It is a flowchart which shows the content evaluation value calculation / extraction process performed by step S453 of FIG. It is a flowchart which shows the preference degree calculation process performed by step S456 of FIG. It is explanatory drawing which shows the content currently broadcast by three channels in a certain time slot | zone regarding one Example of the preference degree calculation apparatus shown in FIG. 1, another Example, and the further another Example. FIG. 12A is an explanatory diagram showing an electronic program guide of “content A”, and FIG. 12B shows an electronic program guide of “content B”, regarding the embodiment of the preference degree calculating apparatus shown in FIG. FIG. 12C is an explanatory diagram showing an electronic program guide for the “content C”. FIG. 13A is an explanatory diagram showing content meta information of “content A” regarding one embodiment, another embodiment, and still another embodiment of the preference degree calculation apparatus shown in FIG. ) Is an explanatory diagram showing the content meta information of the “content B”, and FIG. 13C is an explanatory diagram showing the content meta information of the “content C”. FIG. 7 is an explanatory diagram showing recognition operation of character data and voice data included in “content C” in the character data recognition unit and the voice data recognition unit in the embodiment of the preference degree calculation apparatus shown in FIG. 1. It is explanatory drawing which shows the preference information of "user X" regarding one Example of the preference degree calculation apparatus shown in FIG. It is explanatory drawing which shows the preference degree calculation formula which calculates the preference degree with respect to the extracted keyword regarding one Example of the preference degree calculation apparatus shown in FIG. 1, another Example, and another Example. It is explanatory drawing which shows a mode that the preference degree about a keyword is updated regarding one Example of the preference degree calculation apparatus shown in FIG. It is explanatory drawing which shows a mode at the time of displaying the keyword and the preference degree about the said keyword on a display apparatus regarding one Example of the preference degree calculation apparatus shown in FIG. It is explanatory drawing which shows the fashion degree information which shows the fashion degree of the world about a keyword regarding the other Example of the preference degree calculation apparatus shown in FIG. It is explanatory drawing which shows the content meta-information of "contents F" accumulate | stored in the content storage part regarding the further another Example of the preference degree calculation apparatus shown in FIG. It is explanatory drawing which shows a mode when the server provided with the display apparatus and the preference calculation apparatus and the keyword management apparatus are connected with the network regarding the further another Example of the preference calculation apparatus shown in FIG. It is a block diagram which shows the functional structure of the fashion point calculation apparatus which concerns on one embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Content 12 Electronic program guide 13 Trend information 14 Audience rating information 15 Preference information 100 Preference level calculation apparatus 103 Audience rating input part (Evaluation value calculation means, audience rating input means)
104 fashion level calculation unit (evaluation value calculation means, fashion level calculation means)
105 Preliminary preference calculation unit (evaluation value calculation means, preliminary preference calculation means)
106 Character data recognition unit (character data recognition means)
107 Voice data recognition unit (voice data recognition means)
108 Keyword extraction unit (keyword extraction means)
109 Keyword preference calculation unit (keyword preference calculation means)
150 content recommendation device 152 keyword storage unit (keyword storage means)
153 Keyword display processing unit (display means)
154 Recommended content determination unit (recommended content determination means)
200 Content Recording Output Device 202 Content Recording Processing Unit (Content Recording Unit)
204 Content output processing unit (content output means)
250 Preference level calculation server 251 Viewing history information input unit (viewing history information input means)

Claims (18)

Keyword extracting means for extracting a keyword from content information indicating information about the first content;
An evaluation value calculating means for calculating an evaluation value related to a preference for the second content distributed when the first content and the first content are viewed;
Keyword preference calculation that sets a higher preference for a keyword extracted by the keyword extraction means as the difference between the evaluation value of the first content and the highest evaluation value of the second content is larger A preference degree calculating device.   The preference content calculation apparatus according to claim 1, wherein the content information is an electronic program guide indicating details of content to be distributed. The first content holds content meta information as information related to the content,
The preference content calculation apparatus according to claim 1, wherein the content information is the content meta information. Character data recognition means for recognizing character data in the video information of the first content,
The preference content calculation apparatus according to claim 1, wherein the content information is character data obtained from the character data recognition means. Voice data recognition means for recognizing voice data included in the first content and converting it into text data;
The preference content calculation apparatus according to claim 1, wherein the content information is text data obtained from the voice data recognition unit. Preliminary preference level calculating means for calculating the user's preference level for each of the first content and the second content,
The evaluation value related to the preference for the content is the preference level for each content obtained from the preliminary preference level calculation means,
The keyword preference level calculating means has a larger difference between the preference level of the first content whose preference level is calculated by the preliminary preference level calculation unit and the maximum preference level of the second content level of preference, The preference degree calculation apparatus according to claim 1, wherein a preference degree for a keyword related to the first content is set high . The keyword preference level calculation means increases the difference between the preference level of the first content and the highest preference level among the preference levels of the second content having a higher preference level than the first content. The preference level calculation device according to claim 6, wherein the preference level for keywords related to content is set high . For each of the first content and the second content, a trend level calculation means for calculating a trend level for each content is provided,
The evaluation value related to the preference for the content is a fashion level for each content obtained from the fashion level calculation means,
The keyword preference degree calculating means increases the difference between the popularity degree of the first content whose degree of popularity is calculated by the popularity degree calculating means and the highest degree of popularity of the second content. The preference degree calculation apparatus according to claim 1, wherein a preference degree for a keyword related to one content is set high . The keyword preference level calculating means increases the difference between the trend level of the first content and the maximum trend level among the trend levels of the second content having a higher trend level than the first content. 9. The preference level calculation apparatus according to claim 8, wherein a preference level for keywords related to content is set high . An audience rating input means for inputting an audience rating for each of the first content and the second content;
The evaluation value relating to the preference for content is the audience rating for each content obtained from the audience rating input means,
The keyword preference level calculation means increases the difference between the audience rating of the first content whose audience rating is input by the audience rating input means and the highest audience rating among the audience ratings of the second content. The preference degree calculation apparatus according to claim 1, wherein a preference degree for a keyword related to one content is set high . The keyword preference level calculating means, as the difference between the highest viewing rate of the viewing rate of the second content with the rating and a high audience rating than the first content of the first content is large, the first The preference degree calculation apparatus according to claim 10, wherein a preference degree for a keyword related to content is set high . A preference degree calculating device according to any one of claims 1 to 11, comprising:
A keyword storage means for storing the preference degree of the keyword calculated by the keyword preference degree calculation means and the keyword;
Content recommendation further comprising recommended content determination means for determining a user's preference level for content based on information stored in the keyword storage means and determining recommended content for the user apparatus.   13. The content recommendation device according to claim 12, further comprising a display unit that displays a preference level for the keyword calculated by the keyword preference level calculation unit. A preference degree calculating device according to any one of claims 1 to 11, comprising:
Content recording means for recording content;
A content recording / output device further comprising content output means for outputting the content recorded by the content recording means. A preference degree calculating device according to any one of claims 1 to 11, comprising:
Furthermore, a viewing history information input means for inputting viewing history information by the user is provided,
The preference level calculation server, wherein the keyword extraction unit regards content included in the viewing history information input by the viewing history information input unit as first content. A keyword extraction step of extracting a keyword from content information indicating information about the first content;
An evaluation value calculating step of calculating an evaluation value related to the preference for the first content and the second content broadcast when the first content is being viewed;
Keyword preference calculation that sets a higher preference for the keyword extracted in the keyword extraction step as the difference between the evaluation value of the first content and the highest evaluation value of the second content is larger A preference degree calculating method comprising: a step.   A program for operating the preference degree calculation apparatus according to any one of claims 1 to 11, wherein the preference degree calculation program causes a computer to function as each of the above-described means.   The computer-readable recording medium which recorded the preference calculation program of Claim 17.

Images