JP2003203088A - Programmable contents providing method and system, transmission terminal, program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To effectively provide the related information by collecting the related information in which an end user may be interested in future by estimating the information on end user's future preference during a time when the end user views the contents. <P>SOLUTION: A terminal 60 of the end user mixes a user preference keyword created on the basis of end user's past content viewing history and a keyword of the contents which the end user views at the present, by a certain ratio to create the blended user preference keyword and to update the user preference keyword. Then the contents retrieval event on the basis of the updated user preference keyword is transmitted to SIONet 36, and the related information of the contents is collected. Then the programmable content is created by combining the collected related information and the contents which the end user views at the present, to be provided to the end user. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a condition provided by an end user who desires to view content that meets a certain specific condition using a semantic information network, from the contents provided by a content provider. The present invention relates to a programmable content providing system and a providing method for creating a programmable content by obtaining the content that matches the above and combining the obtained content.

[0002]

2. Description of the Related Art In recent years, a personalized information proposing service has been attracting attention, in which contents that match a user's taste are selected from a vast amount of contents, customized, and proposed. However, in the conventional content providing methods such as advertisements, since the advertisements and related information are provided to the user by using the user preference information extracted based on the meta information attached to the content being viewed by the user and the history of the meta information. On the other hand, only the same advertisement was always provided, and it was not possible to provide relevant information or advertisement that newly cultivates the interest of the user.
In addition, for end users, information such as advertisements presented while watching content is always the same and is often already known information. In most cases, end users are not interested in advertisements. Often, it only interfered with the viewing of content.

[0003]

In the above-mentioned conventional programmable content providing system, advertisement and related information are provided by utilizing user preference information created based on the past content viewing history of the end user. Therefore, there is a problem in that only the same advertisement and related information are always presented, and effective advertisement and related information cannot be presented.

An object of the present invention is to provide an advertisement or related information which the end user may be interested in by estimating the future preference information of the end user while watching the content. Thus, it is to realize a programmable content providing system and method capable of effectively presenting related information and advertisement.

[0005]

In order to achieve the above object, a programmable content providing method of the present invention includes a transmitting terminal for transmitting data as an event and selectively receiving the data transmitted as an event. To
An end user's terminal uses a semantic information network including a receiving terminal to which a filter including an event type and an acquisition condition is set, and the end user's terminal determines the user preference based on the condition of the content desired by the end user. By creating a keyword and transmitting a content search event based on the user preference keyword to the semantic information network, the content matching the condition desired by the end user is obtained from the content provided by the content provider. In the programmable content providing method for providing to the end user, while the end user is watching a certain content, a user preference keyword created based on the past content watching history of the end user and the content currently being watched by the end user. Con Content keyword with a certain ratio to create a blended user preference keyword; and a step of updating the blended user preference keyword with the user preference keyword that serves as a reference for content search. To do.

According to the present invention, the user preference keyword created based on the past content viewing history of the end user and the content keyword of the content currently being viewed by the end user are mixed at a certain ratio and blended user. Since the preference keyword is created and the user preference keyword that serves as a reference for content search is updated, it is possible to search for content that matches the end user's preference.

Further, another programmable content providing method of the present invention comprises a transmitting terminal for transmitting data as an event, and an event type and an acquisition condition for selectively receiving the data transmitted as the event. Using the semantic information network consisting of the receiving terminal to which the filter
By creating a user preference keyword based on the condition of the content desired to be viewed by the end user and transmitting a content search event based on the user preference keyword to the semantic information network, the content provided by the content provider can be changed. A programmable content providing method for obtaining content that meets an end user's desired conditions from among them and providing the content to the end user, wherein the content provider terminal filters content attributes that are attributes of the content being provided. And the user's preference keyword created based on the past content viewing history of the end user while the end user's terminal is watching a certain content, and the end user is currently watching the content. Con Creating a blended user preference keyword by mixing it with a certain content keyword at a certain ratio, and the end user's terminal updating the user preference keyword as a reference for content search by the blended user preference keyword. When,
Providing a content search event to the semantic information network for the end user's terminal to search for information related to content that the end user is currently viewing based on the updated user preference keyword. The content provider terminal, in which the attribute of the content being set as the filter acquisition condition, receives a content search event that matches the filter acquisition condition from the semantic information network, and the content from the end user The step of the content provider terminal receiving the search event returning a response message indicating that the content search event has been received, including the content information that is the information of the provided content, and the response from the content provider terminal. Messe The end user's terminal receiving the message extracts the content information from the received response message, and selects one of the extracted content information that matches the set user preference condition; A step of transmitting a content delivery instruction for requesting the transmission of the content data of the selected content to the content provider terminal that has transmitted the response message, and the content provider terminal receiving the content delivery instruction from the end user terminal Then, the step of delivering the content data requested by the received content delivery instruction to the terminal of the end user, the terminal of the end user receiving the content data from the terminal of the content provider, and the content and the end user Content being viewed And a step of providing the end user to create a programmable content by combining.

According to the present invention, the user preference keyword created based on the past content viewing history of the end user and the content keyword of the content currently being viewed by the end user are mixed at a certain ratio to blend user preference. By creating a keyword, searching related information based on the blended user preference keyword, and combining it with the content that the end user is watching, providing related information that the end user may be interested in the future, It is possible to effectively present related information.

Further, in another programmable content providing method of the present invention, the evaluation information of the end user is obtained based on the operation content of the end user while the terminal of the end user is displaying the content, and the evaluation information is obtained. Based on the blended user preference keywords, a step of updating the blending ratio, which is the ratio of the keywords of the content currently being viewed, may be further included.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION Since the present invention is premised on the use of a semantic information network system which is a distributed network system in which the message nature of information to be transmitted is enhanced, this semantic information network will first be described below. Regarding this semantic information network, Japanese Patent Laid-Open No. 2001-249873 and Takanari Hoshiai et al.
Semantic Information Network Architecture ”, IEICE Transactions B, Vol.J84-B, No.3, pp.411-424 (2001-
3).) Is described in detail. Therefore, only the outline of this semantic information network will be described below.

First, the semantic information network (Semantic I
An nformation-Oriented Network (hereinafter referred to as SIONet) will be described. FIG. 1 shows a network model of SIONet.
Here, for convenience of description, the terminal 2 will be referred to as a transmission terminal 21 of an event sender and a reception terminal 22 of an event receiver. The event receiver uses the receiving terminal 22 to register the semantic information of the event that he / she wants to receive (type of event to be received and acquisition condition) in SIONet1 as meta information. Filter this
Call. On the other hand, the event sender sends an event to SIONet1 by using the transmission terminal 21,
Incentive is given to IONet. As shown in FIG. 2, this event is composed of semantic information describing the characteristics of the event and data. The definition of the semantic information is shown in FIG. The semantic information is the meta information of the event, and
It is an instance of the semantic information type (event type).

SIONet1 is an autonomous decentralized matching network for matching (filtering) the events sent by the event sender against the filter registered by the event receiver. As a result of the matching, the filter that the event passed (reacted to the event) is fired (Igni
Then, the receiving terminal 22 of the corresponding event receiver is autonomously activated. With this mechanism, the target terminal 2 can be searched and discovered in a scalable and real-time manner from an unspecified number of terminals 2.

Next, the event type will be described.
FIG. 4 shows a definition example of an event type which is a template of an event. As shown in FIG. 4, the event type includes an event type name (Event type name), a condition name (in FIG. 4, “Service” and “CPU power” correspond), and a data type for each condition name ( St
ring and Long correspond) and conditional expressions (== and> = correspond) are defined. The event type name is a name for uniquely identifying the event type.

The parent type of the event type can be inherited.

As shown in FIG. 5, an event is created according to the event type data structure. An event is an event type name, a combination of condition name and condition value, and
It is composed of the data part. An error will occur if the condition name, condition expression, or condition value defined in the event does not match the event type. However, the condition name used in the event may be a subset of the event type.

FIG. 6 shows an example of filter definition. The filter uses the event type name (Event type nam
e), attribute name (in Fig. 6, "CPU power" and "Ag"
It consists of a pair of "e") and an attribute value (corresponding to 200 or 25 in FIG. 6). Only the events belonging to the event type defined by the accepted event type name are to be filtered. Can have multiple event type names defined, and
It is also possible to target all events by specifying a wild card (*. *). If the attribute name defined by the filter does not exist in the event type condition name defined by the accepted event type name, an error occurs. However, it may be a subset of event types.

Next, the configuration of SIONet1 will be described.
FIG. 7 is a diagram showing the configuration of SIONet1. As shown in FIG. 7, SIONet1 is a semantic information switch (Sema
nticInformation-Switch, shown as SI-SW in the drawing), Semantic Information-Route
r, shown as SI-R in the drawing), Semantic Information-Gateway (SI-R in the drawing)
(Shown as GW).

The semantic information switch (SI-SW) is a switching mechanism for collating the semantic information registered as a filter with the semantic information given to the event, and as a result, activating the terminal 2 of the event receiver who has fired. provide. The semantic information switch (SI-SW) and each terminal 2 are connected in a star type.

The semantic information router (SI-R) performs event path selection between the semantic information switches and also the terminal 2
It is responsible for transferring the event sent from the to the semantic information switch to another semantic information switch. This is achieved by dynamic event routing based on semantic information.

The semantic information gateway (SI-GW) is
Transfers events between event places. Here, the event place is a minimum unit (ontology domain) that guarantees a common semantic information space. In the event place, the uniqueness of the ontology system such as the name, concept, vocabulary, meaning, and association of the event type is guaranteed, and the semantic information is described based on the common ontology. Basically, the event sent from the terminal 2 of the event sender circulates only in the event place, but the semantic information gateway (SI-
GW) enables mutual distribution of events between event places having different ontology systems. At this time, the semantic information gateway (SI-G
W) performs the ontology conversion of the event and then transfers the event to a different event place.

Next, before explaining the present embodiment,
First, the terms used in this embodiment will be defined. (1) Programmable content providing service collects information related to a content such as a moving image while watching the content, and automatically selects one of the related information collected at this time that matches the user's taste. , At the same time providing it to users at the same time as content such as videos,
Furthermore, the advertisement related to the related information provided in this step is collected, and from the collected advertisements, the one that matches the user's taste is automatically selected, and the advertisement is selected by the user at the same time as the content such as video and related information. It refers to the content combination providing service provided to. In the following description, the combination of contents provided at this time will be referred to as programmable contents or “Prog
(2) Programmable movie refers to a group of video images that combine moving image content such as multiple video clips. For example, movies, dramas, variety programs, music video clip collections, etc. The video program provided to the end user corresponds to “ProgM” or “PM” in the following description, but they are synonymous with each other. A material for a programmable movie, specifically, a video clip, a sound clip, a text, a still image, etc. (4) Link content provides moving picture content to a user in a programmable content providing service. At the same time, the content related to the video provided at the same time (for example, Text content or still image content)
Or, further refers to content related to related content. (5) Content list is SIONet under certain conditions
A group of contents retrieved and collected by using, etc., and a group of contents keywords of those contents. (6) Content attribute refers to information added to content. In the embodiment described below, the content attribute includes a scene type, a cast actor, an age condition, and the like. (7) Content keywords are CA, UA, ProgC creation A because the content attribute added to the content is used in the programmable content providing service.
PL, ProgC reception APL converted into a format that can be commonly understood. By collating this content keyword with the user preference keyword, it is determined whether or not it matches the end user's preference. In the embodiment described below, the content keyword includes a scene type, a cast actor, an age condition, and the like. (8) The scene type represents the characteristics of a scene or a cut image, and means a horror scene, a serious scene, a comedy scene, a love scene, a sensual scene, a landscape video scene, or the like. (9) The user preference attribute (user information) is information indicating the characteristics of the user. User preference attribute (user information)
Is composed of user attributes (age, gender, occupation), user environment attributes (terminal type, time), and service attributes (information indicating service preference). (10) The user preference keyword is information representing the end user's preference for selecting a scene or a cut, and the user information is UA, ProgC creation APL, P
It is converted into a format that can be commonly understood by the logC reception APL. By collating this information with the content keyword, it is determined whether or not the scene or cut matches the end user's preference. In the embodiment described below, the user preference keyword includes a scene type request (simple scene request), an actor request, an age, and the like. (11) Scene hope represents a scene type (preferred) that the end user wants to see, which means a horror scene, a serious scene, a comedy scene, a love scene, a sensual scene, a landscape video scene, and the like, and will be described below. In the embodiment, the desirability for each scene type is given as a weight value. There is also a weight value for the overall scene desire, which indicates how much importance should be placed on the scene type desire. (12) Actor hope represents an actor who the end user wants to see (favorite), and has an actor name as a value. In the embodiment described below, the degree of hope for each actor is given as a weight value. . There is also a weight value for all actor hopes, which indicates how much importance should be placed on actor preference. (13) Large items of user preference keywords are individual items, although the configurations of user preference keywords are roughly classified. In the embodiment described below, scene preference, actor preference, age, etc. correspond to the user preference keyword major items. (14) User preference keyword sub-items are items obtained by further subdividing the individual items (scene desire, actor desire, age, etc.) that make up the user preference keyword. For example, when subdividing a scene desire, which is a major user preference keyword item, horror scene preference, serious scene preference, comedy scene preference, etc. correspond to the user preference keyword minor item. (15) The user preference detailed information refers to a user preference keyword small item. (16) Blend UPK (blend user preference keyword) is a certain ratio of UPK created based on the past content viewing history of the end user and CK (content keyword) of the content currently being viewed by the end user. It is a UPK created by mixing. The UPK created by this mixing is for searching for content that further matches the user preference in relation to the content that the end user is currently watching. (17) Blending ratio of blending UPK is UPK created based on the past content viewing history of the end user.
And the CK of the content that the end user is currently watching,
The latter (CK of content being viewed) occupies the blended UPK created by mixing (% display).
The blend ratio R _B is in the range of 0 ≦ R _B ≦ 200, and R _B =
0 indicates that the blend UPK is entirely composed of past content viewing history, and R _B = 100 indicates the blend UPK.
Indicates that all consist of CK of the content currently being viewed. Further, in the case of R _B > 100, it indicates a UPK that estimates the interest of the future user. (18) The event means a message transmitted to the semantic information network in order to discover information or an information provider that matches one's own preference, using the semantic information network described above. In the embodiment described below, the user preference keyword is inserted into the event. (19) The filter means a message receiving condition set in the semantic information network in order to find information or an information provider that matches one's own preference by using the semantic information network described above. In the embodiment described below, the content attribute is set as a condition in the filter.

Next, the system configuration of a specific embodiment of the present invention will be described. As shown in FIG. 8, a programmable content providing system according to an embodiment of the present invention includes CAs 35 _{1 to} 35 ₃ and content registration APL 34 _1.
And to 34 _3, and ProgC create APL33, ProgC
Reception APL31, UA32, SIONet36,
It is composed of a network 37. The functions of these system components (entities) will be described below.

The CAs 35 _{1 to} 35 ₃ manage the content and the content attributes as agents of the content provider (CP), and deliver the content to the end user using the network 37. In addition, CA35 _{1 to} 35 ₃
Converts the content attribute into a content keyword and presents it to the ProgC creation APL 33 or the like. further,
CA35 ₁ ~35 _3, create a filter based on the content attribute, it is registered in the SIONet36. When the event matching the filter is received from SIONet 36, the content keyword is returned based on the destination information described in the data part of the event.

Content registration APLs 34 _{1 to} 34 ₃ are C
G for P to register the content to CA34 ₁ ~34 ₃
Provide each UI.

The ProgC-created APL 33 is used by an end user or a third-party service provider
Using t36 or the like, the content that is the material of ProgC is searched, and the searched content is combined. Specifically, the ProgC creation APL 33 acquires a blend UPK that fuses the content keyword of the content such as a video currently being watched by the end user and the content viewing history of the end user so far via the ProgC reception APL 31. Create a related content search event from the contents of the blend UPK, and set the event to SIONet.
Call 36. In addition, ProgC created APL33,
Responses from the CAs that have received the event among the CAs 35 _{1 to} 35 ₃ are collected, the content keyword group included in the response message is transmitted to the UA 32 via the ProgC reception APL 31, and the optimal content selection for the UA 32 is performed. Request. The notification of the selected content is received from the UA 32, and the CA that manages the content among the CAs 35 _{1 to} 35 ₃ is instructed to deliver the content determined by the UA 32.

The ProgC reception APL 31 provides a GUI for the end user to view ProgC. Specifically, the ProgC reception APL 31 notifies the ProgC creation APL 33 of this condition when the end user inputs a condition such as a moving image to be watched or a user profile, and provides the moving image content and related content (that is, content combination). Ask. Pro
The gC reception APL 31 notifies the UA 32 of the list of moving image contents and the content (link content) group list related to the content at any time from the ProgC creation APL 33, and requests the UA 32 to narrow down the content to one. Then, the selection result of the UA 32 and the transmission of the content are requested. ProgC receiving AP at the same time
L31 is the blend U presented by UA32 at this time.
The PK is returned to the ProgC creation APL 33, and further requested to collect the related content for this content.
Further, ProgC receiving APL31 is content UA32 is selected, at any time received from the CA that manages the content of the CA35 ₁ to 35 _3, is displayed on the terminal screen of the end user. And ProgC reception APL31
Converts the operation content of the end user during content display into evaluation information for the end user content, and
Notify 32.

The UA 32 manages user preference attributes as an agent of the end user. When the UA32 receives the content keyword list from the ProgC creation APL33 via the ProgC reception APL31, the UA32 calculates the matching degree with the user preference, selects the content to which the content keyword with the highest matching degree is given, and selects the ProgC receiving APL32. Reply to ProgC creation APL33 via. Further, at this time, the UA 32 creates a blend UPK that fuses the content keyword of the selected content and the content usage history of the end user so far, and at the same time, creates the ProgC creation AP via the ProgC reception APL 31.
Reply to L33. When the ProgC reception APL 31 notifies the UA 32 of the evaluation information of the end user obtained based on the operation content of the selected end user who is displaying the content, the UA 32 receives the weight value of the user preference attribute based on the information, or is watching the content. The blend ratio when the content keyword of the content and the content usage history of the user are fused (when the blend UPK is created) is updated.

The SIONet 36 is a network that provides a mechanism for discovering an unspecified number of content providers or contents based on semantic information (meta information), and a user (mainly an end user) of the ProgC-created APL 33 is a ProgC. Used in discovering content that is the source of

It should be noted that SIONet3 in this system
The role of 6 is to provide the function of searching in real time the content that matches a certain condition.
As an alternative entity of t36, a WWW search service or an OMG trading service can be applied. The network 37 is a medium for each entity to perform information communication.

Next, an operation scenario of the entire programmable content providing service system realized by these entities will be described. FIG. 10 shows a method of implementing a programmable content providing service without an intermediary or an intermediary server according to the present embodiment. In this embodiment, a semantic information network is used as Japanese Patent Application No. 2000-0.
It is realized using the event place shown in 62447 and Japanese Patent Application No. 2000-284076. Incidentally, FIG.
Due to space limitations, CP2, CP3, and content registration / control APLs 34 ₂ and 34 ₃ are omitted.

Programmable content provider (CP1
~ 3) and programmable content users (end users) respectively use terminals 60, 61 such as PCs to be used.
In _{1-61 3,} and the network OS of the ORB like CORBA-compliant, No. 2000-062447 and No. 200
It is assumed that the event place factory generation mechanism shown in 0-284076 is installed and connected to the SIONet 36 which is a semantic information network.

The operation scenario of the entire programmable content providing service system can be roughly divided into four steps.

(1) Content registration CP _{1 to 3} are content registration / control APLs 34 _{1 to} 34 ₃
Start each, and register and control this content APL3
4 ₁ ~34 ₃ GUI screen from the content provided by the enter the file name and content attributes of the (video and text, files, such as still images). Content registration / control APLs 34 _{1 to} 34 ₃ use the input information as CA 35 ₁
Register to ~ 35 ₃ . The CAs 35 _{1 to} 35 ₃ create a filter based on the content attribute, and register the filter for receiving the content search event in the SIONet 36.

(2) ProgC request The end user activates the ProgC reception APL 31 and inputs the profile of the end user and information of a favorite moving image. At this time, the ProgC reception APL 31 discovers the ProgC creation APL 33 existing on the network 37 or the ProgC installed APL 33 is installed on the same terminal.
The C creation APL 33 is activated to request the creation and provision of ProgC.

(3) ProgC creation (content combination
Let's ProgC creation APL33 is ProgC receiving APL3
Video search event from the video request information notified from 1.
Create and match the event using SIONet36 etc.
Collect the content group to be used. ProgC created APL3
3 is connected to UA32 via ProgC reception APL31.
Requested the selection of Tents and blended with the content selection result
Get UPK. Regarding the selected content,
CA35 ₁~ 35₃C that manages the relevant content
Instruct A to deliver to ProgC reception APL31,
Content related to content from Blend UPK
Create a search event and use IONet36 etc.
Collect the content group that matches the event.

Thereafter, in the same sequence, the ProgC creation APL 33 transmits the UA3 via the ProgC reception APL 31.
2 requests the content selection, obtains the content selection result and blend UPK, and instructs the CA managing the content of CAs 35 _{1 to} 35 ₃ to the ProgC reception APL 31 regarding the selected content. Then, the process of creating a content search event related to the content from the blend UPK and collecting the content group that matches the event using the SIONet 36 or the like is repeated. (4) ProgC appreciation ProgC received APL31 displays the content distributed from the CA35 ₁ to 35 ₃ in APL screen, the end user while viewing the displayed contents, "rewind (re Listening)", "stop ( Operations such as "end, close)", "screen enlargement", and "screen reduction" are performed. Upon receiving this operation input, the ProgC reception APL 31 converts the operation content into an evaluation value and feeds it back to the UA 32.

Next, FIG. 11 shows an example of the components and functional configuration of the programmable content providing system. In the programmable content providing system, the UA factory 38 and the CA factory 3 are added to the configuration shown in FIG.
9 is newly shown.

The ProgC reception APL 31 has functions for GUI, moving picture display, moving picture stream reception, and UA generation request. The ProgC creation APL 33 has a function for performing interface, link setting, and link content collection. Content registration / control APL34
Has a function for GUI, content attribute input navigation, and CA generation request. UA32 is API,
It has functions for user preference attribute management, attribute → keyword creation, link selection, and history collection. CA35 is
It has functions for API, moving image delivery, attribute → keyword creation, and content attribute management.

Next, each block constituting the entity shown in FIG. 14 will be described.

Content registration / control application 3
As shown in FIG. 12, 4 includes a CA generation request information transmission unit 41, CA generation information 42, a CP authentication information input screen display unit 43, a CP authentication information input unit 44, and a content attribute registration unit 45. , And a content registration section 49.

The CP authentication information input screen display section 43 displays the CP
The CP authentication information input unit 44 notifies the CA generation information creation unit 42 of the authentication information input by the CP. CA generation information creation unit 42
Creates CP generation request information from the CP authentication information notified from the CP authentication information input unit 44. The CA generation request information transmission unit 41 transmits the CA generation request information to the CA factory 39.

The content registration section 49 comprises a content registration screen display section 46, a content attribute input section 47, and a content preview display section 48.

The content registration screen display unit 46 displays a screen for the CP to register the content, the content preview display unit 48 displays the preview screen of the content being registered, and the content attribute input unit 47 inputs it by the CP. The created content attribute is registered in the content attribute registration unit 4
Notify 5. The content attribute registration unit 45 registers the content attribute in CA.

As shown in FIG. 13, the CA factory 39 has a CP authentication information management unit 54 and a CP authentication unit 55.
And a CP authentication information receiving unit 56 and a CA generating unit 57.

The CP authentication information receiving section 54 receives the CP authentication information from the content registration / control APL 34. CP
The authentication unit 55 is the C notified from the CP authentication information receiving unit 56.
The P authentication information is collated with the CP authentication information managed by the CP authentication information management unit 54. CA generation unit 57 is CA
To generate. The CP authentication information management unit 54 manages CP authentication information.

As shown in FIG. 14, the CA 35 includes a content search event reception filter registration unit 71, a content search event reception filter creation unit 72, a content attribute / control information reception unit 73, and a content search event reception unit 74. And a content search response transmission unit 76,
The content keyword creating unit 77, the content attribute managing unit 78, the instruction content extracting unit 79, the content delivery instruction receiving unit 80, and the content delivering unit 81.

The content attribute / control information receiving section 73
A content attribute or content control information is received from the content registration / control APL 34. The content search event reception filter creation unit 72 creates a filter for receiving a content search event from the content attribute. The content search event reception filter registration unit 71 sets the content search event reception filter to SI.
Register with ONet36.

The content search event receiving section 74 receives an event (content search event) that matches the content search event reception filter. The content search response transmission unit 76 returns a content keyword to the ProgC creator who transmitted the content search event.

The content keyword creating unit 77 creates a content keyword from the content attribute. The content attribute management unit 78 manages content attributes.
The content delivery instruction receiving unit 80 uses the ProgC creation AP.
The content delivery instruction from L33 is received. The instruction content extraction unit 79 extracts the content delivery instruction content received from the ProgC creation APL 33. The content delivery unit 81 delivers the content to the designated content delivery destination.

ProgC creation application (Pro
The gC creation APL) 33, as shown in FIG. 15, is a content request receiving unit 82 and a content delivery instructing unit 83.
And a content search unit 84 and a content search result distribution unit 89.

The content request receiving unit 82 uses the ProgC
A content (moving picture, text, advertisement, etc.) request message from the reception APL 31 is received, and the requested content information is notified to the content delivery instruction unit 83. The content delivery instruction unit 83 delivers the content (video, text, advertisement, etc.) requested by the ProgC reception APL 31 to the ProgC reception APL 31 which is the request source.
A content delivery instruction message is transmitted to the CA 35 that manages the content.

The content retrieving unit 84 searches the SIONet 36 for content that meets the conditions requested by the end user.
, Etc., and search for related content with respect to the content provided to the end user.
Search using 36, etc. Furthermore, the content search unit 8
4 is a blend UPK extraction unit 85, a content search event creation unit 86, and a content search event transmission unit 8
7 and a content search response unit 88.

The blend UPK extraction unit 85 is a ProgC
When a content request or a first video search request is received from the reception APL 31, a blend UPK indicating information serving as a keyword for searching video content or related content to the content is extracted and notified to the content search event creation unit 86. To do. The content search event creation unit 86 uses the blend UPK extraction unit 8
A content search event is created from the blend UPK notified from 5. Content search event transmission unit 87
Sends the content search event created by the content search event creation unit 86 to SIONet 36. The content search response receiving unit 88 receives the response message from the CA 35 that has received the content search event.

The content search result distribution unit 89 uses the CA3.
The content keywords included in the response message from 5 are extracted, all the received content keywords are notified to the ProgC reception APL 31 as a content list, and the selection of the content is requested.

ProgC receiving application (Pro
The gC reception APL) 31 is, as shown in FIG.
A generation request information transmission unit 105 and UA generation information creation unit 1
06, the initial setting unit 107, and the moving image search requesting unit 110
A ProgC reception requesting unit 111, a content selection requesting unit 118, a content requesting unit 119, a content information list receiving unit 120, and a content receiving unit 121.
A content display control unit 123 and an evaluation information creation unit 1
26 and an evaluation information notification unit 127.

The initial setting section 107 is composed of a user information input screen display section 108 and a user information input section 109. The user information input screen display unit 108 displays
A screen for inputting user information is displayed, and the user information input unit 109 receives the user information input by the end user.

The UA generation information creating section 106 creates UA generation information from user information. The UA generation request information transmission unit 105 transmits the UA generation information created by the UA generation information creation unit 106 to the UA factory 38.

In addition, the ProgC reception request unit 111 includes
The moving image search condition input unit 112 is provided, and the moving image search condition input unit 112 presents a screen for inputting conditions such as a title or genre of a moving image or a video program that the end user wants to watch, a performing actor, or the like.

The moving picture search request unit 110 notifies the ProgC creation APL 33 of the moving picture search conditions, and requests ProgC creation and provision. Content information list receiving unit 12
0 receives the list of link content information from the ProgC creation APL 33, and notifies the content selection requesting unit 118 of this content information list. Since the content selection requesting unit 118 requests the UA 32 to select one content from the content information list, the content selection requesting unit 118 notifies the UA 32 of this content information list and requests the response (content selection result and blend UPK) for the content. Notify the department 119. Content request unit 11
9 should deliver the content selected by UA32.
The ProgC creation APL 33 is requested and the acquired blend UPK is also notified to the ProgC creation APL 33.

The content receiving section 121 receives the content distributed from the CA 35.

The content display control unit 123 includes a moving image display unit 113, a moving image operation input unit 114, a text display unit 115, a text operation input unit 116, a still image (advertisement) display unit 117, and a still image (advertisement). ) Operation input unit 122
It consists of and.

The moving picture display section 113 displays a moving picture, and the moving picture operation input section 114 accepts an input (skip, cue, repeated reproduction, stop, size enlargement, size reduction, etc.) input to the movie being watched by the end user. , The content of the operation is notified to the content requesting unit 119 and the evaluation information creating unit 126.

The text display unit 115 displays text, and the text operation input unit 116 operates the text that the end user is viewing (size enlargement, size reduction,
(Display off, etc.) Input is accepted, and the operation content is notified to the evaluation information creation unit 126.

The still image display unit 117 displays a still image such as an advertisement, and the still image operation input unit 122 operates the still image being viewed by the end user (size enlargement, size reduction,
(Display off, etc.) Input is accepted, and the operation content is notified to the evaluation information creation unit 126.

The evaluation information creating unit 126 creates evaluation information from the operation contents of the end user. Evaluation information notification unit 12
7 notifies the UA 32 of the evaluation information created by the evaluation information creating unit 126.

The UA factory 38, as shown in FIG. 17, includes a user authentication information management section 130 and a user authentication section 13.
1, a user authentication information reception unit 132, and a UA generation unit 13
3 and 3.

The user authentication information receiving section 132 has a Prog
The user authentication information is received from the C reception APL 31. The user authentication unit 131 compares the user authentication information notified from the user authentication information receiving unit 132 with the user authentication information managed by the user authentication information management unit 130. UA
The generation unit 133 generates the UA32. The user authentication information management unit 130 manages user authentication information.

As shown in FIG. 18, the UA 32 includes a content list receiving section 134 and a content selecting section 13
5, a user preference keyword / content keyword matching unit 136, a user preference keyword creating unit 137,
A user attribute receiving unit 138, an environment attribute collecting unit 139,
Preference information receiving unit 140 and user preference attribute management unit 141
And a user preference attribute value learning unit 142, a content evaluation receiving unit 143, and a blend UPK creating unit 144.

The user attribute receiver 138 receives the user profile information of the end user. The environment attribute collection unit 139 is for collecting environment information (CPU capacity, resolution, HD
Capacity, line capacity, etc.). Preference information receiver 140
Receives service-specific preference information (program genre, actors, etc.). The user preference attribute management unit 141 manages user preference attributes (user attributes, environment attributes, preference information, respective weight values, service usage history, etc.). The content list receiving unit 134 outputs the content list linked to the content currently being viewed to the ProgC receiving AP.
It receives from L31, and notifies it to the content selection part 135.

The content selection unit 135 notifies the content list to the user preference keyword / content keyword collation unit 136 in order to select the content from the content list notified by the content reception unit 134, acquires the collation result, and retrieves the content. elect. User preference keyword / content keyword matching unit 136
Compares the user preference keyword with the content keyword of each content shown in the content list, and notifies the content selection unit 135 of the matching result. The user preference keyword creation unit 137 creates a user preference keyword from the user preference attribute.

The content evaluation receiving unit 143 uses the Prog
C Receive content evaluation information from APL31,
The content evaluation information is provided to the user preference attribute value learning unit 14
Notify 2. The user preference attribute value learning unit 142 updates the user preference attribute value using the content evaluation information from the content evaluation receiving unit 143.

The cooperation between the entities including the above functional blocks will be described in detail with reference to the following implementation example.
First, I will explain the operation scenario in the functional block, and then
A user preference attribute learning method in the UA 32 will be described.

The processing procedure when the end user views the programmable content will be described below.

(1-1) Here, a case where the CP ₁ providing the moving image content generates the CA 35 ₁ will be described. CP1 is its terminal 61 ₁ to the content registration and control applications 34 ₁ and CA factories (C
AF) 39 is installed and the content registration / control application 34 ₁ is started.

(1-2) The content registration / control application 34 ₁ activates the CAF 39, issues an operation for generating CA 35 ₁ from the API provided by the CAF 39, and generates CA 35 ₁ .

(1-3) The content registration / control application 34 ₁ displays a content registration window as shown in FIG. In the content registration screen shown in FIG. 19, the content ID is automatically input and the file name is input by the explorer. Further, the scene type is selected from the options, and the user can select a plurality of options. Performing actors are manually input, and it is possible to select from the options whether the performing actors are the leading role, the semi-leading role, or the leading role. In addition, the age condition is input by a numerical value, and the video title is manually input.

(1-4) CP1 registers the content as a video clip file such as an MPEG file from the content registration window, and the content attribute (scene type, cast actor, age condition of the user for viewing) in the content. Etc.) respectively. In addition,
The registered content (that is, scene) is automatically identified by the content registration unit 49 (content ID).
Is given. The information entered is the content registration /
Content attribute registration unit 4 of control application 34 ₁
5 is notified, and the CA 35 ₁ generated in the above process is registered.

(1-5) The content attribute / control information receiving section 73 in the CA 35 ₁ notifies the content search event reception filter creating section 72 of the content attribute of the content registered as described above. When the content search event reception filter creation unit 72 creates a filter as shown in FIG. 21 by the method shown in FIG. 20, the content reception filter registration unit 71 creates the filter as SIONet.
Register at 36. The filter of FIG. 21 is an example of a filter created from the content attributes input as shown in FIG. In FIG. 21, $ xxx indicates the property value of the event property whose property name in the event is xxx. Also, character strings are enclosed in single quotation marks "', and wild cards are indicated by *. For example, 'Black *' refers to all character strings starting with'Black ', and' * Black * 'is'Black'.
Refers to all existing strings. Further, this filter is for the content provider that provides the moving image content to receive the moving image content search event.

(1-6) CPs 2 and 3 that provide contents (texts and still images) other than moving images are also the same as in (1-1) above.
The CAs 35 ₂ and 35 ₃ are generated in the same manner as in the steps (1-5) to register contents on the screens shown in FIGS. 22 and 23. CAs 35 _{1 to} 35 ₃ convert the registered content attributes into the filters shown in FIGS.
Register with ONet36. $ Acto in Figure 25
r == “*” means that anyone can do it.

(2-1) The end user (ProgC receiver) uses the ProgC receiving application 31, the UA factory (UAF) 39, and the ProgC on his or her terminal 60.
Install the creation application 33 and
The C reception application 31 is activated.

(2-2) The ProgC reception application 31 activates the UAF 39, issues an operation for generating the UA 32 from the API provided by the UAF 39, and generates the UA 32.

(2-3) The ProgC reception application 31 displays the user profile registration window as shown in FIG.

(2-4) The user of the ProgC reception application 31 receives the user profile such as address and age, authentication information, service preference information (favorite movie genre, actor, etc.) from this user profile registration window. Input to the information input unit 109.

(2-5) The user information input unit 109 transmits the input authentication information to the generated UA 32 to perform user authentication, and then the user information input unit 109
The input user preference attribute is transmitted to the UA 32 to manage the user preference attribute.

(2-6) Next, in order to start using the programmable content, the end user wants to watch a moving image from the content search menu window screen as shown in FIG. 27 presented by the ProgC reception request unit 111 of the ProgC reception APL 31. When the content conditions are input and “video content search” is selected, the video search request unit 110
The ProgC creation APL 33 is activated, and the input moving image content request condition is notified to the ProgC creation APL 33.

(3-1) In the ProgC creation APL 33, the content search event creation unit 86 creates a video content search event as shown in FIG. 28 using the user's video request condition and the user profile notified by the ProgC reception APL 31. The content search event transmission unit 87 creates the event and transmits the event to the SIONet 36.

(3-2) The content search event receiving section 74 of the CA 35 ₁ which has registered the filter as shown in FIG. 21 in the SIONet 36 receives the content search event. The content search event reception unit 74 extracts a reply destination (event sender, that is, the object reference of the ProgC creation APL 33) from the received content search event, and notifies the content search response transmission unit 76 of it.

(3-3) Content search response transmitting section 76
Acquires a content keyword from the content keyword creating unit 77, creates a reply message as shown in FIG. 29, and sends it back to the above reply destination. The reply message shown in FIG. 29 is a reply message to the content search, and includes content that is a candidate for the material of ProgC.

(3-4) ProgC creation The content search response reception unit 88 of the APL 33 collects all reply messages from the CA 35 and notifies the content search result distribution unit 89 of the collected content keywords. The content search result distribution unit 89 distributes the above list of content keywords to the ProgC reception APL 31.

(3-5) ProgC reception When the content information list receiving unit 120 of the APL 31 receives the content information (content keyword) list from the content search result distribution unit 89, the content information list receiving unit 120 selects from the content information list for the end user. In order to select one of the most appropriate contents, the selection menu screen as shown in FIG. 30 is presented to the end user for selection, or the UA is selected.
It is necessary to have 32 automatically select. Here, UA32
This will be explained using a case where the user is requested to select. When the content information list receiving unit 120 notifies the content selection request unit 118 of the received content information list, the content selection request unit 118 delivers the list to the UA 32.

(4-1) Upon receiving the content information list from the content selection requesting unit 118, the content list receiving unit 134 of the UA 32 receives the content selection unit 135.
Notify the received content information list to. The content selection unit 135 uses the content information list as a user preference keyword (UPK) / content keyword (C
K) Notify the collation unit 136 and request the collation score calculation between UPK and all CKs.

(4-2) The UPK / CK collating unit 136
Upon receiving the collation score request from the content selection unit 135, the user preference keyword (UPK) generation unit 137 receives a U
Request PK creation. The UPK creation unit 137 integrates the profile and preference information input by the user himself, and the content usage history of the user (input from the content evaluation receiving unit) up to now to construct a keyword, and at the same time, selects a major item (user UPK as shown in FIG. 31, FIG. 32, or FIG. 33 is created by using the value of W _k obtained by the following equation (1) as the weight value of the category of preference attribute).

FIG. 31 is a diagram showing the content keywords used in ProgC, and FIG. 32 shows the structure of user preference keywords for matching with the content keywords in FIG.

[0094]

[Equation 1]

Here, X and Y are learning parameters,
ω _k is the change rate of the weight value, and t indicates the time.

UPK is user preference attribute information (consisting of multi-attributes, each of which is given weight information) acquired from the user preference attribute management unit 141, that is, the age value of the end user, and the end user frequently appreciates it. It is composed of information such as movie genre, cast actor name and its weight value (small item weight value), and large item weight value representing the scene desirability and actor desirability. The former (small items) is shown in Figure 31,
The scene type name S ₁ in FIG. 32 or 33,
S _2, ···, S _m and weight values _{w S1, w S2, ···,} w
_Sm , actor names A ₁ , A ₂ , ..., A _n and weight values w _A1 ,
, etc. _{w A2, ···, w A n} , age value Age. The latter is a scene desirability weight value wS, an actor desirability weight value WA, and the like. The method of collecting and learning these values will be described later.

(4-3) The UPK / CK collating unit 136
As shown in FIG. 34, a collation score calculation formula is created using the above UPK. FIG. 35 shows a specific example of creating this collation score calculation formula. 34 and 35, the profile is an indispensable condition and is not related to the score.

34 and 35, scene type name S _U1 , scene type name S _U2 , ... _Are values of character strings, for example, S _U1 = “comedy scene”, S _U2 = “action scene”. “And so on. Further, the actor name A _U1 , the actor name A _U2 , ... _Are values of character strings, for example, A _U1 = “Jackie Chan”, A _U2 = “Tom Cruise” and the like.
Furthermore, $ xxx is the content keyword,
As for the small item, it is a weight value of the small item whose item name is xxx. For example, in the content keyword of FIG.
$ S ₁ is V _s1 and $ A ₁ is V _A1 .

All CKs are applied to this collation score calculation formula, and each collation score is calculated as shown in FIG. In the example of FIG. 36, the content A is in the content information list.
The content B has been selected because the content B has a high collation score. The UPK creation unit 137 assigns an ID to the UPK created at this time, and outputs the time t corresponding to the ID.
n and the values of Xk and Yk at this time are held. UP
The K / CK collation unit 136 compares this content selection result with U.
The PK of the PK is notified to the content selection unit 135.

(4-4) The content selection unit 135 also blends the CKs of the selected contents with the UPK creation unit 14.
4 and request to create a blend UPK. The blend UPK creating unit 144 creates a blend UPK by blending CK and UPK of the selected content at a certain ratio by the method shown in FIG. 37, and notifies the content selecting unit 135 of the blend UPK.

(4-5) The content selection unit 135 outputs the above content selection result and blend UPK to ProgC.
Reply to the received APL 31.

(5-1) The content selection requesting unit 118 of the ProgC reception APL 31 has the content selecting unit 135.
Upon receiving the content selection result and the blend UPK from the content request unit 119, the content request unit 119 is notified of them, and the content request unit 119 notifies them of the ProgC creation APL33.
Notify to.

(5-2) The content request receiving unit 82 of the ProgC creation APL 33 receives the ProgC reception APL 31.
When the content selection result and the blend UPK notified from the content request unit 119 are received, the blend UPK
Is notified to the blend UPK extraction unit 85, and the selected content is notified to the content delivery instruction unit 83.

(5-3) The content delivery instruction section 83
The CA 35 that manages the selected content is instructed to deliver the content to the ProgC reception APL 31. Specifically, the object reference and content ID of the ProgC reception APL 31 are notified.

(5-4) When the content delivery instruction receiving unit 80 of the CA 35 receives the content delivery instruction from the content delivery instruction unit 83 of the ProgC creation APL 33, the instruction content extraction unit 79 instructs the delivery destination ProgC receiving APL 31 to send it. The object reference is extracted, the content file corresponding to the content ID is read, and the content delivery unit 81 is notified.

(5-5) The content delivery section 81 delivers the above-mentioned content file to the ProgC reception APL 31 which can be referred to by the object reference.

(5-6) On the other hand, ProgC created APL3
The blend UPK extraction unit 85 of 3
The PK is notified to the content search event creation unit 86.

(5-7) The content search event creation unit 86 uses the blend UPK notified from the blend UPK extraction unit 85 in a method as shown in FIG.
A content (text information) search event as shown in 9 is created, and the content search event sending section 87 sends this content search event to the SIONet 36. The text information search event shown in FIG. 39 is
It is created from the blended UPK using the CK of the moving image content. After that, from (3-2) to (5-
By performing the same processing as 6), the text content is searched for instead of the video content, one text content is selected from the text content, and the blend UPK is acquired.

(5-8) At the same time as the text content distribution instruction, the still image (advertisement) content is searched and selected by the same processing from (3-2) to (5-6). In this process, an advertisement search event as shown in FIG. 40 is created and used as the content search event. This advertisement search event is created from the blended UPK using the CK of the text information content.

(6-1) Upon receiving the content distributed by the CA 35, the content receiving section 121 of the ProgC receiving application 31 notifies the content display control section 123.

(6-2) Content display control section 123
Notifies the moving image display unit 113 if the content notified from the content receiving unit 121 is a moving image file, the text display unit 115 if the content is text content, or the still image display unit 117 if the content is still image content. The display unit displays the content as shown in FIG.

(6-3) The end user views the contents such as moving images in the display windows displayed by the respective display units, and "maximizes the screen", "rewind", depending on the end user's preference. Click the "Skip" button. Alternatively, it is possible to watch the movie until the end without any operation.

(6-4) When the operation button is clicked by the end user, or when the content such as the moving image currently being viewed is completely viewed (immediately before the end of viewing),
From the content display control unit 123 to the evaluation information creation unit 126
Is notified of the operation content, the content ID, the content keyword, and the ID of the related UPK. The evaluation information notification unit 127 notifies the UA 32 of such information (evaluation information). The method of converting the operation information into the evaluation value and the relationship with UPK small item learning will be described later.

(7-1) In the above step (3-1), the retrieval of a moving image by the ProgC-created APL 33 is a trigger for content combination in this service, and is a prerequisite. Therefore, the provision of this service to the end user is ended by the end user clicking the “End” button in FIG. 41 or by the ProgC creation APL 33 ending the moving image search. It is possible to combine the system of Japanese Patent Application No. 2001-246603 "Programmable Movie Service" as a method for the ProgC creation APL 33 to automatically retrieve the next moving image. In this case, not only is the content combination such as appropriate text information and still images (advertisements) automatically performed according to the end user's preference, but the main content video content is also expanded according to the user preference (video content The connection relationship will change.

In this service, the UA 32 handles the multiple user preference attribute shown in FIG. Then U
A method of determining (learning) multi-attribute weight information of the user preference attribute information of A32 will be described.

The weight values of the major items in FIG. 33, the actor desirability weight value W1 (= WA), and the scene desirability weight value W2
(= WS), and the learning method will be described.

Let each weight value be Wk, (k = 1, 2, ...
.), And the range of the weight value is 0 ≦ Wk ≦ 100.
Letting n be the number of times of learning, t be the UA usage time (elapsed time) of the end user and the client application, etc., and ωk be the amount of change per unit time when changing the weight value in trial and error. ) Is given by the calculation formula. Here, the values of Xkn and Ykn fluctuate as the learning progresses, as in equations (2) and (3).

[0118]

[Equation 2]

[0119]

[Equation 3]

In the above equations (2) and (3), V
n is an evaluation value extracted from the behavior of the end user, and the behavior of the end user and the evaluation value have a correspondence relationship as shown in FIG. Also,

[0121]

[Outer 1]

As shown in FIG. 44, t is the value of t when the content Cn that is the factor of the feedback is specified (selected), and is the respective values of Xk and Yk when t = tn. . Here, the evaluation value V is −1 ≦ V ≦ 1
If V> 0, it means that appropriate content selection has succeeded, and the user is satisfied, and if V <0, then appropriate content selection fails. However, it indicates that the user is not satisfied.
In addition, Tkn is a value calculated by the following formula (4), where tn is a value of t when the user preference keyword used when the content causing the behavior of the end user is specified is created. Is.

[0123]

[Equation 4]

By feeding back this evaluation, the parameters Xkn and Ykn of the function for calculating the weight value Wk are updated. However, the range is 0 ≦ Xkn ≦ π
/ 2, -π / 2 ≦ Ykn ≦ π / 2. When it exceeds the range, it is corrected as shown in the following formula (5).

[0125]

[Equation 5]

Next, the determination (learning) of the multi-attribute weight information of the UA 32 will be described.

In the large item weight value learning calculation formula of the UA32 shown in the above formula (1), the following parameters are set to Pro
The gC reception APL 31 or the like sets the UA 32.

As examples of user preference attributes, two actor preferences and scene type preferences are handled, and in the above equation (1), k = 1 is the actor preference and k = 2 is the scene type preference.
That is, W1 is the weight value of the actor preference attribute, and W2 is the weight value of the scene type preference attribute.

As initial values of the weights, W1 = 50 and W2 = 5
Set to 0. That is, X10 = 0, Y10 = 0, X20 = 0,
Set Y20 = 0.

In order to make the amount of change in the weight value per unit time in trial and error different from each other, ω1 = 2.
0 and ω2 = 1.2.

Under the above parameter settings, the value of the actor preference attribute (favorite actor name) and the value of the scene type preference attribute (favorite scene type) are extracted from the movie watching history of the end user so far. Assuming that the value is set by the above, the UA 32 learns the weight value of each attribute as follows.

[0132] The end user views a certain programmable content, and some time when the end user performs some operation, the evaluation based on the operation is performed over time.
Extracted by C reception APL31 and its evaluation value V is U
It is fed back to A32. When the end user sees the content selected by the UA 32 to the end, V = 0.1 is given, but it is generally considered that this evaluation often occurs. However, when a scene that does not match the end user's interest appears, it may be skipped by the end user, and V =-
Considering that the evaluation of 0.5 is given, it is assumed that the end user's evaluation value Vn is created by creating the user preference keyword at time tn and using the user preference keyword as shown in FIG. At that time, the weight values W1 and W2 of the user preference keywords to be evaluated are also calculated from the above equation (1) and shown in the table of FIG. In this example, it is assumed that there is a scene that has little relation to the main story in the early stage of the movie and is not of interest to the end user, and there is an operation such as fast forward or frame advance. In this simulation, t is the elapsed time (unit: minutes) of watching the movie. The learning results are shown in the graphs of FIGS. 46 and 47. FIG. 46 shows the learning result of the weight value of W1 or the actor preference attribute, and FIG. 47 shows the learning result of W2 or the weight value of the scene type preference attribute.
From this result, the end user has a stronger actor preference than a scene type preference, and the weight values learned by the UA32 converge on W1 in the range of 40 to 50 and W2 in the range of 10 to 20, respectively. I know I'm going. In addition, in the weight value learning calculation formula by the above-described formula (1), by setting different amounts of change in the weight value per unit time in trial and error, the UA 32 allows the UA 32 to perform one type of action performed by the end user. It is possible to simultaneously learn the weight value of each attribute from the feedback information of.

Next, a method of setting the weight value of the small item shown in FIG. 33 will be described.

As shown in FIGS. 48 and 49, the weight value of the small item is updated when the end user views the content. n indicates the number of evaluations. Vn is the end user's evaluation of the content selected using the nth user preference keyword (UPK), and γ is the number of times the end user has watched ProgC in the last month. As shown in the following equation (6), the coefficient μ to the value to be added (or subtracted) from the weight value of the small item is calculated from γ. Further, in FIG. 49, actor names A _c1 , A _c2 , ... Are values of character strings, for example, “Tom Cruz”, “Keanu Lee Bass”.
It becomes a value like. Further, the actor names A _U1 , A _U2 , ... Are values of character strings, for example, values such as “Tom Cruz” and “Blood beat”.

[0135]

[Equation 6]

FIG. 48 shows a quantitative value (weight) indicating how much each scene type such as a love scene, a comedy scene, an action scene, and a horror scene is preferred with respect to the end user's scene type preference (ie program genre). (Value) is extracted from the content keyword of the viewed content and reflected in the weight value. FIG. 48
The content keyword and the user preference keyword in FIG. 3 indicate only the information portion related to the scene type, and other attributes are omitted for the sake of space.
In the content keyword, the scene type name S _c1 ,
S _c2 , ... Is a character string such as “love scene” or “comedy scene”, and weight values w _c (S _c1 ), w _c (S _c2 ), ... Correspond. For example, the content (value) of S _c1 is “love scene” (hereinafter S _c1 =
“Love scene”) and w _c (S _c1 ) = 8
If it is 0, the love scene degree of the content (a score representing the degree of being liked by end users who like the love scene on a 100-point scale) is 80. On the other hand, in the user preference keyword, scene type names S _U1 , S
_U2 ... is a character string such as “action scene” or “love scene”, and weight values w _U (S _U1 ), w _U (S _U2 ), ... Correspond to each scene type. To do. For example, S _U1 = “action scene”, and w _U (S _U1 ).
= 50, the score representing the action scene preference of the end user with 100 points is 50.

Here, a method of feeding back the content keyword to the user preference keyword of the end user will be described. Extract all combinations of k and i such that S _Uk = S _cj . That is, small items having the same scene type name in both keywords are extracted. Then, for the sub-item, the weight value w _U (S _Uk ) is updated by a calculation formula such as the following formula (7).

[0138]

[Equation 7]

However, the value range of w _U (S _Uk ) is 0 ≦ w _U (S
_{Since Uk} ) ≦ 100, if the value is out of the range, it is corrected as in the following equations (8) and (9).

[0140]

[Equation 8]

[0141]

[Equation 9]

In the equation (8), when the maximum weight value 100 is exceeded, the exceeded amount (Δw _U (S _Uk ) in FIG. 48)
Not only subtracting the weight value w _U (S _Uk ) of the relevant scene type, but also the weight value w of another scene type name
_U (S _U1 ), w _U (S _U2 ), ... W _U (S _Ui ), ...
The value of is also subtracted by Δw _U (S _Uk ) as shown in the following expression (10).

[0143]

[Equation 10]

If the value range is exceeded even in this subtraction, the weight value of the scene type is set to 0 as in the case of the equation (9).

FIG. 49 is a quantitative value (weight value) showing how much the end user likes each of actors such as "Tom Cruz", "Keanu Livas", and "Jackie Jean" with respect to the end user's actor preference. Is extracted from the content keyword of the viewed content and reflected in the weight value. In FIG. 49, each of the content keyword and the user preference keyword shows only the information part regarding the actor, and other attributes are omitted for the sake of space. In the content keyword, the actor names A _c1 , A _c2 , ... Are character strings such as “Keanu Livas” and “Tom Cruz”, and weight values w _c (A _c1 ), w _c (for each actor name). A _c2 ), ... Corresponds.
For example, the content (value) of A _c1 is “Keanuly bus” (hereinafter referred to as A _c1 = “Keanuly bus”), and w
_{Assuming that c} (A _c1 ) = 80, the Keanuly bass degree of the content (a score expressing the degree of being liked by the end user who likes the Keanuly bass on a 100-point scale) is 80. On the other hand, in the user preference keyword, the actor names A _U1 , A _U2 , ... Are character strings such as “Keanu Lee Bass” and “Jackie Jean”, and the weight value w _U (A _U1 ), W _U (A _U2 ), ... Corresponds. For example, A _U1 = “Keanuly Bus”,
Assuming that w _U (A _U1 ) = 50, the end user's Keanuly bass preference is expressed in 100 points, and the score is 50.

Here, the user preference key of the end user
To feed back content keywords to the code
Will be explained. A_Uk= A_cjAll k and i combinations
To extract. That is, the actor name is
Extract the matching sub-items. Then, regarding the sub-item
Then, the weight value w can be calculated by the following formula (11). _U
(A_Uk) Is updated.

[0147]

[Equation 11]

However, the range of w _U (A _Uk ) is 0 ≦ w _U (A
_{Since Uk} ) ≦ 100, if the value exceeds the range, it is corrected as shown in equations (12) and (13).

[0149]

[Equation 12]

[0150]

[Equation 13]

In the equation (12), when the maximum weight value 100 is exceeded, the exceeded amount (Δw in FIG. 49).
_U (A _Uk)) only well subtracts the weight value w _U (A _Uk) with respect to the actor, the weight value w _U (A to other actors
The values of _U1 ), w _U (A _U2 ), ..., W _U (A _{U i} ), ... Are also subtracted by Δw _U (A _Uk ) as shown in equation (14).

[0152]

[Equation 14]

If the value range is exceeded even in this subtraction, the weight value for the actor is set to 0 as in the case of the equation (13).

Next, the method of determining the blend ratio and the learning method necessary for creating the blend UPK of the UA will be described.

FIG. 37 shows an example in which the content viewing history of the end user so far is blended with the CK of the content h being viewed by the end user. In the example of FIG. 37, the CK ratio (R _B ) is 75%.
However, in the present embodiment, the blend ratio R _B takes a value range of 0 ≦ R _B ≦ 200, and R _B = 0.
Indicates that the blend UPK consists entirely of history, R
_B = 100, the blend UPK consists of CK. R _B > 1
In the case of 00, the blend UPK is obtained by removing the elements included in the end user history from the currently viewed CK.

A method of learning the blend ratio R _B will be described with reference to the following equations (15) to (19). The number of times of learning n, end users and UA usage time, such as the client application (elapsed time) t, the amount of change per unit time when changing the weight values as omega _R in trial and error, the blending ratio R _B of the following It is given by equation (15).

[0157]

[Equation 15]

Here, XRn and YRn are expressed by the following formula (1
The value fluctuates as learning progresses as in 6) and equation (17).

[0159]

[Equation 16]

[0160]

[Equation 17]

In the above equations (16) and (17), Vn is an evaluation value extracted from the behavior of the end user, and the behavior and the evaluation value of the end user are the same as those in the large item weight value learning method. The correspondence relationship is as 43. Also,
Tn is defined by the following equation (18), where tn is the value of t when the blend UPK that is the basis of the event in which the content that caused the behavior of the end user is searched is created.
It is a value calculated as follows.

[0162]

[Equation 18]

The parameters XRn and YRn of the function for calculating the blend ratio R _B are updated by feeding back the above evaluation. However, each range is 0 ≦ X _RN ≦
π / 2 and −π / 2 ≦ Y _RN ≦ π / 2. If it exceeds the range, it is corrected as in the following equation (19).

[0164]

[Formula 19]

Next, a method of creating the blend UPK will be described. As described below, a blend ratio R of CK (content keyword) and UPK (user preference keyword)
Blending with _B means that the CK of the content being viewed is R _B
(%), It is to create a UPK set subsections weight value of the blend UPK history of the end user in the ratio of 100-R _B (%). Sub-item A _Um in Blend UPK
The weight value w _B (A _Um ) of is calculated by the following formula (20).

[0166]

[Equation 20]

However, the range of w _B (A _Um ) is -100.
Since ≦ w _B (A _Um ) ≦ 100, if the value exceeds the range, it is corrected as shown in the following formulas (21) and (22).

[0168]

[Equation 21]

[0169]

[Equation 22]

In the equation (21), when the maximum value 100 of the weight values is exceeded, not only the weight value w _B (A _Um ) for the actor is subtracted by the exceeded amount (Δw _B (A _Um )), Weight values for other actors w _B (A _U1 ), w _B (A _U2 ), ...
.., w _B (A _Ui ), ... Are also subtracted by Δw _B (A _Um ) as shown in the above equation (22). When the value exceeds the range in this subtraction, the weight value for the actor is set to −100 as in Expression (23).

[0171]

[Equation 23]

As described in the previous section, UPK and CK are data consisting of a finite number of items and weight values for those items, and can be treated as vectors in a multidimensional space. Therefore, the above-mentioned history records UPK and CK are shown in FIG. UA is a content search UPK created by mixing history UPK and CK at a certain ratio.
Present to the application as K (Blend UPK). The content search UPK (blend UPK) created at this time exists on the line segment AB like the point X in FIG.

Here, the ratio of the history UPK and CK in the content search UPK (blend UPK) and their characteristics will be described. If the ratio of CK is R _B %, the history UPK
The ratio is 100-R _B%. When R _B is close to 0, a blend UPK that emphasizes the user's past interest is created, and when the value of R _B is near 100, the blend UPK that emphasizes the user's current interest is created. We also targeted the range of 100 or more as the range of R _B handled by UA. R _B > 10
In the case of 0, the content search UPK (blend UPK) created by the UA is the straight line A as shown by the point Y in FIG.
It exists on B and is on an extension of point B when viewed from point A. Point A
Can be said to be the future because point B is the present and point B is the present. Therefore, blending UP with R _B > 100
Infer the user's future preferences by creating K,
It is possible to actively develop the interest of the user. To summarize the relationship of R _B and blend UPK following Table 1
become that way.

[0174]

[Table 1]

The blend UPK creation function by the UA described above is simulated assuming the ProgC service. In this simulation, Prog
As an example of the C service, a service that provides a combination of a video clip content, text information related to an actor who appears in the video clip, and an advertisement image of groceries, talent goods, etc. We used the "Actor" attribute and the "Food" attribute for food and drink preferences. At this time, what kind of content search UPK (blend UPK) the UA creates in ProgC is shown below.

When the text content indicated by CK as shown in FIG. 51 (b) is delivered to the user having the past history as shown in FIG. 51 (a), the advertisement image as the third content is displayed. CK ratio R _B as UPK to be searched
Various characteristics can be created by. A typical generation example is shown below. (1) R _B = 20% (CK 20%, history UPK 80
%) Is shown in FIG. 51 (c). (2) A blend UPK created with R _B = 200% (CK is 200% and history UPK is −100%) is shown in FIG.
Shown in.

In the case of using the former blend UPK, FIG.
As in 1 (c), the history so far is emphasized, and it is considered unlikely that a new content that the user has not viewed in the past is proposed. On the other hand, when the latter blend UPK is used, as shown in FIG. 51 (d), the weight value for a small item that the user has seen so far is very small, and a new value that the user has never had. It can be said that there is a high possibility of proposing content.

In the present embodiment described above, SIONe is used.
Although the method of realizing the present system by using t has been described, the WWW search service can be replaced with SIONet. When using the WWW search service,
4, the content search event reception filter registration unit 71 and the content search event transmission unit 87 in FIG. 15 are respectively used as a directory service (WWW search service)
The content registration message sending unit and the content search message sending unit have been replaced with a keyword indicating an actor name, a scene type, etc. as a search keyword, and the content registration and the search message are sent to HTTP (Hyper)
This can be realized by transmitting with the Text Transfer Protocol).

As described above, according to the programmable content providing system of the present embodiment, while the end user is watching a certain content, an advertisement or a relation which the end user may have an interest in the future can be obtained from the content. Information can be provided to cultivate new interests of users. Also, information such as advertisements presented by the end user while watching the content is associated with the end user from an unexpected connection, and the end user often becomes interested in these advertisements, etc. Content can be enjoyed even more. Furthermore, by combining the system of Japanese Patent Application No. 2001-246603 "Programmable Movie Service", not only is the content combination such as appropriate text information and still image (advertisement) automatically performed according to the end user's taste, The moving image content, which is the main content, is also a service whose expansion (connection relationship of moving image contents) changes according to user preference.

In the programmable content providing system of this embodiment, the text information is used as the related information to be combined with the content such as the moving image, and the advertisement related to the text information is combined with the content such as the moving image. However, advertisements may also be included in this related information.

Although not shown in the figure, in the programmable content providing system of this embodiment, the terminal 60 of the end user and the content providers 1-3.
Terminals 61 ₁ to 61 ₃ is provided with a recording medium recording a program for executing the programmable content providing method described above. This recording medium may be a magnetic disk, a semiconductor memory or another recording medium. This program is executed from the recording medium to the terminal 60 of the end user by the ProgC reception APL31 and the ProgC creation A.
Loaded as PL33, content providers 1 to
Content point registration / control APL on ₃ terminals 61 _{1 to} 61 ₃
Read as 34 _{1 to} 34 ₃ and end user's terminal 6
0 and the terminals 61 _{1 to} 6 of the content providers 1 to 3
Control the operation of 1 ₃ . Specifically, the terminal 60 of the end user and the terminals 61 _{1 to} 6 of the content providers 1 to 3
1 ₃ in the CPU above processing realized by instructing to perform specific processing hardware resources of the terminal 61 ₁ to 61 ₃ of the terminal 60 and the content provider 1-3 of the end user under control of the program To be done.

[0182]

As described above, according to the present invention,
A user preference keyword created based on the past content viewing history of the end user and a content keyword of the content currently being viewed by the end user are mixed at a certain ratio to create a blended user preference keyword, and the blended user preference Providing advertisements or related information that may be of interest to the end user in the future by searching for advertisements or related information based on keywords and combining it with the content that the end user is viewing, The effect that an effective presentation can be performed can be acquired.

[Brief description of drawings]

FIG. 1 is a diagram showing a model of a semantic information network.

FIG. 2 is an explanatory diagram showing the structure of an event.

FIG. 3 is an explanatory diagram showing the definition of semantic information.

FIG. 4 is an explanatory diagram showing a definition example of an event type.

FIG. 5 is an explanatory diagram showing an example of an event.

FIG. 6 is an explanatory diagram showing a definition example of a filter.

FIG. 7 is a diagram showing a configuration of a semantic information network.

FIG. 8 is a diagram showing all entities in the programmable content providing service system of the present specification.

FIG. 9 is a diagram showing an example of an operation image of an application and an agent in the programmable content providing service system.

FIG. 10 is a diagram showing an example of an information delivery sequence of an application and an agent in the programmable content providing service system.

FIG. 11 is a diagram showing an example of components and functional configurations of a programmable content providing system.

FIG. 12 is a diagram showing an example of a configuration of a content (movie scene) registration / control application 34.

FIG. 13 is a diagram showing an example of a configuration of a CA factory 39.

FIG. 14 is a diagram showing an example of a configuration of a CA 35.

FIG. 15 is a diagram showing an example of a configuration of a programmable content (ProgC) creation application 33.

16 is a diagram showing an example of a configuration of a programmable content (ProgC) reception application 31. FIG.

FIG. 17 is a diagram showing an example of a configuration of a UA factory 38.

FIG. 18 is a diagram showing an example of a configuration of a UA32.

FIG. 19 is a diagram showing an example of a moving image content registration and content attribute setting screen.

FIG. 20 is a diagram showing an example of a mapping method from content keywords to content search event reception filters.

FIG. 21 is a diagram showing an example of a filter for receiving an event for searching moving image content.

FIG. 22 is a diagram showing an example of a text content registration and content attribute setting screen.

FIG. 23 is a diagram showing an example of a filter for receiving an event for searching text content.

FIG. 24 is a diagram showing an example of a still image (advertisement) content registration and content attribute setting screen.

FIG. 25 is a diagram illustrating an example of a filter for receiving an event for searching for still image (advertisement) content.

FIG. 26 is a diagram showing an example of a profile registration screen of an end user who receives programmable content.

FIG. 27 is a diagram showing an example of a moving image content request condition input screen of an end user who receives programmable content.

FIG. 28 is a diagram showing an example of an event for searching moving image content.

FIG. 29 is a diagram showing an example of a message sent back to the event sender by the CA that has received the ProgC-compatible content search event.

FIG. 30 is a diagram showing an example of a screen on which a list of moving image contents searched by using an event for searching moving image contents is displayed, and an end user selects from the list.

FIG. 31 is a diagram showing an example of content keywords used in matching whether or not the content is suitable for the user preference of the end user.

FIG. 32 is a diagram showing an example of user preference keywords used in matching whether or not the content is suitable for the end user's user preference.

FIG. 33 is a diagram showing an example of a method of finding a matching score between a content keyword and a user preference keyword using a matching score calculation formula and selecting which content is appropriate.

FIG. 34 is a diagram for explaining conversion from a user preference keyword to a matching score calculation formula.

FIG. 35 is a diagram showing an example of a user preference keyword and an example of a matching score calculation formula.

[Fig. 36] Fig. 36 is a diagram illustrating a calculation example of a matching score between a user preference keyword and a content keyword.

FIG. 37 is a diagram showing an example of creating a blended UPK of UA.

[Fig. 38] Fig. 38 is a diagram illustrating an example of a mapping method from a blend UPK to a content search event.

FIG. 39 is a diagram showing an example of an event for searching text contents.

FIG. 40 is a diagram showing an example of an event for searching for still image (advertisement) content.

FIG. 41 is a diagram showing an example of a programmable content viewing screen by an end user.

FIG. 42 is a diagram showing an example of multiple attributes expressing content attributes and user preference attributes.

[Fig. 43] Fig. 43 is a diagram illustrating a correspondence relationship between an end user's behavior and an evaluation value.

FIG. 44 is a diagram illustrating an example of a feedback and learning method for calculating a weight value of a large item of UA.

FIG. 45: User preference keyword creation time and U at that time
It is a figure which shows the example of an end user's evaluation with respect to the content selected using the user preference keyword which A created.

FIG. 46 is a diagram showing an example of a learning result of a large item weight value W1.

FIG. 47 is a diagram showing an example of a learning result of a large item weight value W2.

[Fig. 48] Fig. 48 is a diagram illustrating an example of a method of learning and setting a weight value of a small item of UA-scene type preference.

[Fig. 49] Fig. 49 is a diagram illustrating an example of a method for learning and setting a weight value of a small item of UA-actor preference.

FIG. 50 is a diagram showing an example of vector display of UPK, CK, and blend UPK.

FIG. 51 is a diagram showing an example of creating a blend UPK with different blend ratios.

[Explanation of symbols]

1 Semantic Information Network (SIONet) 2 Terminal 10 Semantic Information Network 21 Transmission Terminal 22 Reception Terminal 31 ProgC Reception APL 32 User Agent (UA) 33 ProgC Creation APL 34 Content Registration / Control APL 35 Content Provider Agent (CA) 36 SIONet 37 Network 38 UA Factory 39 CA Factory 41 CA Generation Request Information Transmitting Section 42 CA Generation Information Creating Section 43 CP Authentication Information Input Screen Display Section 44 CP Authentication Information Input Section 45 Content Attribute Registration Section 46 Content Registration Screen Display Section 47 Content Attribute Input Section 48 content preview display unit 49 content registering unit 54 CP authentication information management unit 55 CP authentication unit 56 CP authentication information receiving unit 57 CA generating unit 60 end-user of the terminal 61 _{1-6 3} of the content provider terminal 71 the content search event receiver filter registration section 72 the content search event receiver filter generation unit 73 content attribute and control information receiving unit 74 content search event receiving unit 76 content search response transmitting unit 77 content keyword generation unit 78 content attribute management Part 79 Instruction content extraction unit 80 Content delivery instruction reception unit 81 Content delivery unit 82 Content request reception unit 83 Content delivery instruction unit 84 Content search unit 85 Blend UPK extraction unit 86 Content search event creation unit 87 Content search event transmission unit 88 Content search Response reception unit 89 Content search result distribution unit 105 UA generation request information transmission unit 106 UA generation information creation unit 107 Initial setting unit 108 User information input screen display unit 109 User Information input unit 110 Video search request unit 111 ProgC reception request unit 112 Video search condition input unit 113 Video display unit 114 Video operation input unit 115 Text display unit 116 Text operation input unit 117 Still image (advertisement) display unit 118 Content selection request unit 119 Content requesting unit 120 Content information receiving unit 121 Content receiving unit 122 Still image (advertisement) operation input unit 123 Content display control unit 130 User authentication information management unit 131 User authentication unit 132 User authentication information receiving unit 133 UA generation unit 134 Content list Reception unit 135 Content selection unit 136 User preference keyword / content keyword matching unit 137 User preference keyword creation unit 138 User attribute reception unit 139 Environmental attribute collection unit 140 Preference information reception unit 141 User preference attribute management unit 42 the user preference attribute value learning unit 143 content evaluation receiving section 144 blends UPK creation unit

Claims (16)

[Claims] 1. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. In a programmable content providing method for providing to an end user content that meets the conditions desired by the end user from among the contents provided by the content provider by transmitting to the network, the end user views the content. Inside the en Creating a blended user preference keyword by mixing the user preference keyword created based on the user's past content viewing history and the content keyword of the content that the end user is currently viewing at a certain ratio; A method of providing programmable content, comprising the step of updating a user preference keyword that serves as a reference for content search with the preference keyword. 2. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. What is claimed is: 1. A programmable content providing method for transmitting to a network to obtain content that matches a condition desired by an end user from among the content provided by the content provider and providing the content to the end user. However, A step of setting a content attribute which is an attribute of Tents as a filter, and a user preference keyword created based on a past content viewing history of the end user while the end user's terminal is viewing a certain content. And a step of creating a blended user preference keyword by mixing the content keyword of the content currently being viewed by the end user with a certain ratio, and the end user's terminal using the blended user preference keyword as a reference for content search. And a content search event for the end user's terminal to search for information related to the content currently being viewed by the end user based on the updated user preference keyword. The semantic information network And a step of receiving, from the semantic information network, a content search event in which the terminal of the content provider in which the attribute of the provided content is set as a filter acquisition condition, receives a content search event that matches the filter acquisition condition. A step in which the terminal of the content provider that has received the content search event from the end user returns a response message indicating that the content search event has been received, including content information that is information on the provided content; A step in which the end user's terminal receiving the response message from the provider's terminal extracts content information from the received response message and selects one of the extracted content information that matches the set user preference condition; , End The step of the user terminal transmitting a content delivery instruction for requesting the transmission of the content data of the selected content to the content provider terminal that has transmitted the response message, and the content provider terminal from the end user terminal. Receiving a content delivery instruction and delivering the content data requested by the received content delivery instruction to an end user's terminal; the end user's terminal receiving the content data from the content provider's terminal; A programmable content providing method, comprising the step of creating programmable content by providing the end user with programmable content by combining the content and the content being watched by the end user. 3. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. A programmable content providing method for obtaining content that matches a condition desired by an end user from among the content provided by a content provider by transmitting the content to a network and providing the content to the end user, wherein the terminal of the end user is provided. But the end user has While watching the content, the user preference keywords created based on the past content viewing history of the end user and the content keywords of the content currently being viewed by the end user are mixed at a certain ratio to create a blended user preference keyword. A step of creating, a step in which the terminal of the end user updates a user preference keyword serving as a reference at the time of content search based on the blended user preference keyword, and a step in which the end user terminal updates the user preference keyword Based on the step of sending a content search event to the semantic information network to search for information related to the content that the end user is currently watching, and the end user's terminal receiving the response message from the content provider's terminal , Received response message Extracting content information from the page, selecting one of the extracted content information that matches the set user preference condition, and the end user terminal requests transmission of content data of the selected content. A content delivery instruction for transmitting to the terminal of the content provider that has transmitted the response message, the end user's terminal receives the content data from the content provider's terminal, and the content and the end user are watching A method of providing programmable content, comprising the step of creating programmable content by combining the content and providing the content to an end user. 4. The end user's evaluation information is obtained based on the operation content of the end user when the end user's terminal is displaying content, and the current one of the blend user preference keywords is calculated based on the evaluation information. The programmable content providing method according to any one of claims 1 to 3, further comprising a step of updating a blend ratio, which is a ratio of a keyword of the content being viewed. 5. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. A program for causing a computer to execute a programmable content providing method of acquiring content that meets an end user's desired condition from among the content provided by a content provider by transmitting to a network and providing the content to the end user. The record that was recorded As a medium, while the end user is watching a certain content, a certain ratio of the user preference keyword created based on the end user's past content viewing history and the content keyword of the content currently being watched by the end user. A record in which a program for causing a computer to execute the processing of mixing to create a blended user preference keyword and the processing of updating the blended user preference keyword to update the user preference keyword that serves as a reference for content search Medium. 6. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. A program for causing a computer to execute a programmable content providing method of acquiring content that meets an end user's desired condition from among the content provided by a content provider by transmitting to a network and providing the content to the end user. The record that was recorded A medium, wherein the terminal of the content provider sets a content attribute which is an attribute of the provided content as a filter, and the end user's terminal is watching the content while the end user is watching the content. A process of mixing a user preference keyword created based on a user's past content viewing history and a content keyword of the content currently being viewed by the end user to create a blended user preference keyword; The terminal updates the user preference keyword that serves as a reference when searching for content using the blended user preference keyword, and the terminal of the end user is currently viewed by the end user based on the updated user preference keyword. Search for information related to content in For transmitting the content search event for the purpose to the semantic information network, and the content provider event in which the attribute of the provided content is set as the acquisition condition of the filter, the content search event that matches the acquisition condition of the filter. Content from the semantic information network, and a content provider terminal that has received a content search event from the end user and has received a response message indicating that the content search event has been received. A process of returning information including information, and the end user terminal receiving the response message from the content provider terminal extracts the content information from the received response message, and the user preference set from the extracted content information. To the condition The process of selecting a match, the process of the end user's terminal transmitting a content delivery instruction for requesting the transmission of the content data of the selected content, to the terminal of the content provider that transmitted the response message, and the content The terminal of the provider receives the content delivery instruction from the terminal of the end user, and delivers the content data requested by the received content delivery instruction to the terminal of the end user. A program for causing a computer to execute processing of receiving content data from a terminal, creating programmable content by combining the content with content being viewed by the end user and providing the content to the end user is recorded. recoding media 7. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. A program for causing a computer to execute a programmable content providing method of acquiring content that meets an end user's desired condition from among the content provided by a content provider by transmitting to a network and providing the content to the end user. The record that was recorded A medium, wherein the end user's terminal is viewing a certain content while the end user is viewing the content, the user preference keyword created based on the past content viewing history of the end user, and the content currently being viewed by the end user. A process of creating a blended user preference keyword by mixing the content keyword with the blended user preference keyword at a certain ratio, and a process of updating the user preference keyword serving as a reference at the time of content search by the terminal of the end user. A process in which the terminal of the end user transmits a content search event for searching information related to content currently being viewed by the end user to the semantic information network based on the updated user preference keyword, Response message from content provider terminal The terminal of the end user who receives the message extracts the content information from the received response message and selects one of the extracted content information that matches the set user preference condition. The process of transmitting a content delivery instruction for requesting the transmission of the content data of the selected content to the terminal of the content provider that transmitted the response message, and the end user's terminal transmitting the content data from the terminal of the content provider. A recording medium having a program recorded thereon for causing a computer to execute a process of receiving a content, creating a programmable content by combining the content with a content being viewed by an end user, and providing the programmable content to the end user. 8. The end user's evaluation information is obtained based on the operation content of the end user when the end user's terminal is displaying content, and the current one of the blend user preference keywords is calculated based on the evaluation information. The recording medium according to any one of claims 5 to 7, wherein a program for further executing a process of updating a blend ratio, which is a ratio of a keyword of content being viewed, is recorded. 9. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. Using the semantic information network, the terminal of the end user creates a user preference keyword based on the condition of the content desired to be viewed by the end user, and a content search event based on the user preference keyword is used as the meaning information. A program for causing a computer to execute a programmable content providing method of acquiring content that meets an end user's desired condition from among the content provided by a content provider by transmitting to a network and providing the content to the end user. And then While the user is watching a certain content, the user preference keyword created based on the past content watching history of the end user and the content keyword of the content currently being watched by the end user are mixed at a certain ratio and blended user A program for causing a computer to execute a process of creating a preference keyword and a process of updating a user preference keyword serving as a reference for content search with the blended user preference keyword. 10. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. End user's terminal using the semantic information network
By creating a user preference keyword based on the condition of the content desired to be viewed by the end user and transmitting a content search event based on the user preference keyword to the semantic information network, the content provided by the content provider can be changed. A program for causing a computer to execute a programmable content providing method of obtaining content that meets an end user's desired conditions from among and providing the content to the end user, which is provided by a terminal of the content provider. A process of setting a content attribute, which is an attribute of content, as a filter, and a terminal of the end user, while the end user is watching a certain content, a user preference key created based on the past content viewing history of the end user. A process of creating a blended user preference keyword by mixing a word and a content keyword of the content currently being viewed by the end user at a certain ratio; and when the end user terminal searches for content using the blended user preference keyword A process of updating a reference user preference keyword, and a content search event for the end user's terminal to search for information related to content currently being viewed by the end user based on the updated user preference keyword. Is transmitted to the semantic information network, and the terminal of the content provider, in which the attribute of the provided content is set as the acquisition condition of the filter, sends a content search event matching the acquisition condition of the filter to the semantic information network. With the process received from A process in which the terminal of the content provider that receives the content search event from the end user returns a response message indicating that the content search event has been received, including the content information that is the information of the provided content; A process in which the end user's terminal receiving the response message from the provider's terminal extracts content information from the received response message and selects one of the extracted content information that matches the set user preference condition. , A process in which the end user's terminal sends a content delivery instruction for requesting the transmission of content data of the selected content to the content provider's terminal that sent the response message, and the content provider's terminal sends the end user's terminal Content from A process of receiving a delivery instruction and delivering the content data requested by the received content delivery instruction to the terminal of the end user, and the terminal of the end user receiving the content data from the terminal of the content provider, A program for causing a computer to execute a process of creating programmable content by providing the end user with the content being watched by the end user. 11. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. End user's terminal using the semantic information network
By creating a user preference keyword based on the condition of the content desired to be viewed by the end user and transmitting a content search event based on the user preference keyword to the semantic information network, the content provided by the content provider can be changed. A program for causing a computer to execute a programmable content providing method of obtaining content that meets an end user's desired conditions from among and providing the content to the end user, wherein the terminal of the end user is an end user. While viewing the content, the user preference keyword created based on the past content viewing history of the end user and the content keyword of the content currently being viewed by the end user are mixed at a certain ratio to blend the user preference key word. A process of creating a code, a process of updating the user preference keyword, which is a reference for content search by the blended user preference keyword, by the end user's terminal, and the end user's terminal updates the user A process of transmitting a content search event for searching information related to content currently being viewed by the end user to the semantic information network based on the preference keyword, and an end user receiving a response message from the terminal of the content provider. The terminal extracts the content information from the received response message and selects one of the extracted content information that matches the set user preference condition, and the end user's terminal selects the content of the selected content. Content delivery for requesting data transmission A process of transmitting an instruction to the terminal of the content provider that has transmitted the response message, and a terminal of the end user receives the content data from the terminal of the content provider, and combines the content with the content being viewed by the end user. A program for causing a computer to execute a process of creating programmable content and providing it to an end user. 12. The end user's evaluation information is obtained based on the operation content of the end user when the end user's terminal is displaying content, and the current one of the blend user preference keywords is calculated based on the evaluation information. The processing for updating the blend ratio, which is the ratio of the keyword of the content being viewed, is further executed.
The program according to any one of 1. 13. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. End user's terminal using the semantic information network
By creating a user preference keyword based on the condition of the content desired to be viewed by the end user and transmitting a content search event based on the user preference keyword to the semantic information network, the content provided by the content provider can be changed. In a programmable content providing system that obtains content that meets an end user's desired conditions from the inside and provides the content to the end user, while the terminal of the end user is watching the content, the end user's past The user preference keyword created based on the content viewing history of the content and the content keyword of the content currently being viewed by the end user are mixed at a certain ratio to create a blended user preference keyword. A programmable content providing system characterized by updating a user preference keyword, which is a reference for content search, by a mode. 14. A transmission terminal that transmits data as an event, and a reception terminal to which a filter including an event type and an acquisition condition is set in order to selectively receive the data transmitted as an event. End user's terminal using the semantic information network
By creating a user preference keyword based on the condition of the content desired to be viewed by the end user and transmitting a content search event based on the user preference keyword to the semantic information network, the content provided by the content provider can be changed. A programmable content providing system that obtains content that meets the end user's desired conditions from the inside and provides the content to the end user, by setting the content attribute that is the attribute of the provided content as a filter and acquiring the filter. When a content search event that matches the conditions is received from the semantic information network, a response message indicating that the content search event has been received is returned including the content information that is the information of the provided content, and the end user When receiving the content delivery instruction from the terminal, the content provider terminal that delivers the content data requested by the received content delivery instruction to the end user's terminal, and the end user while watching certain content. A user preference keyword created based on the user's past content viewing history and a content keyword of the content currently being viewed by the end user are mixed at a certain ratio to create a blended user preference keyword, and the blended user preference keyword is created. According to the above, the user preference keyword that is a reference in the content search is updated, and based on the updated user preference keyword, the content search event for searching the information related to the content that the end user is currently viewing is the meaning. In the information network When sending and receiving a response message from the terminal of the content provider, the content information is extracted from the received response message, and one of the extracted content information that matches the set user preference condition is selected and selected. The content delivery instruction for requesting the transmission of the content data of the content is transmitted to the terminal of the content provider that has transmitted the response message, and the content data from the terminal of the content provider is received,
A programmable content providing system, comprising: an end user's terminal that creates programmable content by providing the end user with programmable content by combining the content and the content being watched by the end user. 15. A semantic information network comprising a plurality of terminals to which a filter including an event type and an acquisition condition is set in order to selectively receive data transmitted as an event, A transmitting terminal that transmits data as an event, and while the end user is watching a certain content, the user preference keyword created based on the past content watching history of the end user and the content currently being watched by the end user. The content keyword of the above is mixed at a certain ratio to create a blended user preference keyword, the blended user preference keyword is used to update the user preference keyword that serves as a reference in content search, and based on the updated user preference keyword , Information related to the content that the end user is currently watching. When a content search event for searching for is sent to the semantic information network and a response message is received from the terminal of the content provider, the content information is extracted from the received response message and set from the extracted content information. Select one that matches the user preference condition, send a content delivery instruction for requesting the transmission of the content data of the selected content, to the terminal of the content provider that has sent the response message, and from the terminal of the content provider. Receive content data,
A transmission terminal that creates programmable content by providing the end user with the content and the content that the end user is watching. 16. The terminal of the end user obtains evaluation information of the end user based on the operation content of the end user when displaying content, and based on the evaluation information, among the blended user preference keywords. The programmable content providing system according to claim 13 or 14, which updates a blend ratio, which is a ratio of the keyword of the content currently being viewed.