JP4211940B2 - Information providing apparatus and information providing method - Google Patents

Description

  The present invention relates to an apparatus and method for providing information via a communication network, and more particularly to an information providing apparatus that distributes information related to content that a user is interested in to a user terminal device.

2. Description of the Related Art Conventionally, content-based filtering and collaborative filtering have been adopted as methods for recommending content in devices that recommend content (digital works) such as information, services, and television programs according to user interests. Of these, content-based filtering characterizes the content, for example, content description information composed of a plurality of keywords, and the user's interest in the content from the user's history of the content, such as directly inputting the browsing status of the content and the suitability of interest. When information is extracted and content is newly provided, a content evaluation value is predicted from the content description information and user interest information, and content with a large evaluation value is provided as recommended content information (for example, patents) References 1 and 2).
JP 2002-142160 A JP 2003-114903 A

  However, in the conventional information providing system, the recommended content for the same user differs depending on the type of terminal used by the user, and the recommended content that accurately reflects the user's interest is not always provided. There's a problem.

  In the conventional information providing system, the content description information is created by an information provider who provides recommended content information to a user or used by an information provider purchasing from a content description information provider who creates content description information. The information provider recommends content adapted to the user's interest from only one piece of content description information. Due to the spread of the Internet, there are many information providers, and many content explanation information providers provide information for one content, and the content explanation information providers enhance the contents explanation information. The degree is different. For example, taking a movie as an example, in Japanese WEB sites, explanation information about Japanese movies is detailed, but explanation information about Western movies is poor. On the other hand, information about movies in the United States is abundant on a WEB site in the United States, whereas information about movies outside the United States may be small. In addition, since the types of terminals possessed by users have become abundant, it is difficult for information providers to provide information to all terminals of all users. For example, information is provided via the Internet using a mobile phone, but information is not provided via VBI (Vertical Blanking Interval) of television broadcasting using a television receiver. As described above, the content explanation information is different for each information provider, and the conventional information providing system has a problem that the result of information recommendation is different for each information provider even for the same user.

  Further, the fact that the information provider varies depending on the terminal owned by the user means that the user's history information used for extracting the user's interest cannot be managed centrally. For example, when a user uses two information providers, user interest information obtained from a user's history at one information provider cannot be transmitted to the other information provider. In other words, even if the same user has the same content description information and the same user history, the user interest information of one information provider is different from the user interest information of the other information provider. Thus, the conventional information providing system also has a problem that the result of information recommendation is different for each information provider.

  In addition, in the case of dramas, even if the user occupies a large specific gravity as an element that shows interest, and in the case of a movie, performers who occupy a large specific gravity as an element that shows interest should use the same recommendation method. In other words, there is a problem that the user cannot select the recommendation method regardless of the content description information.

  Thus, if the information provided by the information terminal is different, information that can be confirmed on one terminal cannot be confirmed on another terminal. For example, information recommended for a television receiver in the morning differs from information recommended for a mobile phone in the daytime and outdoors. As a result, when the recommended program was confirmed on the TV receiver in the morning, but there was no time to make a recording reservation, even if an attempt was made to make a recording reservation using a mobile phone outdoors, If it is different from the television receiver, it is necessary to search for the program again to perform scheduled recording, and the user has to perform troublesome operations.

  Therefore, the present invention has been made in view of the above problems, and an object thereof is to provide an information providing apparatus and the like that can provide recommended content that accurately reflects a user's interest.

  In addition, as a more specific object, the present invention also aims to provide an information providing apparatus and the like that can provide the same recommended content information without depending on a terminal operated by a user or a terminal viewed. To do.

  In order to achieve the above object, an information providing apparatus according to the present invention is an information providing apparatus that provides information related to content that a user is interested in to the terminal device of the user, and describes an explanation regarding a plurality of types of content. Content description information management means for holding a plurality of types of content description information, user interest information management means for holding user interest information that is information relating to the interests of a plurality of users, and a plurality of types of contents held by the content description information management means Refer to the content description information selection means for selecting and reading the content description information used to determine the content that the user is interested in from the content description information, and the user interest information held in the user interest information management means Is read out by the content explanation information selecting means. Based on the content description information, characterized in that it comprises a providing means for providing the recommended content determining means, information relating to the determined recommended content to the terminal device of the user to determine the content to be recommended to the user.

  In order to achieve a more specific object, the information providing apparatus further includes a reception unit that receives access by a user terminal device via the communication network, and the content description information management unit includes a plurality of different types. Holding a plurality of pieces of content description information describing a plurality of pieces of content associated with each type of terminal device, wherein the content description information selection means is based on the type of the terminal device of the accessing user, Recommended content description information used for determining content that the user is interested in from among a plurality of content description information stored by the content description information management means, and providing content description information for providing to the user Select and read, the recommended content determination means, the user interest By referring to the user interest information held in the management means, information relating to the interest of the accessing user is identified, and based on the identified information and the recommended content explanation information read out by the content explanation information selection means The content to be recommended to the user is determined, and the providing means specifies information relating to the determined recommended content from the providing content description information read by the content description information selecting means, and the specified information Is provided to the terminal device of the user.

  The present invention can be realized not only as such an information providing apparatus, but also as an information providing method using processing means provided in the information providing apparatus as a processing step, or a program for causing a computer to execute the method, Alternatively, it can be realized as a computer-readable recording medium in which the program is recorded.

  According to the present invention, recommended content that accurately reflects a user's interest is provided. More specifically, the same recommended content can be easily acquired by the same user regardless of the information provider or the user terminal.

  Further, detailed explanation information can be obtained regardless of the terminal depending on the user's interest. That is, the recommended content that accurately reflects the user's interest is provided without depending on the type of the terminal operated or browsed by the user.

  Therefore, according to the present invention, it is possible to uniformly and quickly obtain recommended content information according to the user's interest regardless of the information provider or the user's terminal, and various electronic devices are inexpensive, highly functional, and diversified. Today, when one person uses many terminals, it can be said that the practical value of the present invention is extremely high.

  An information providing apparatus according to the present invention is an information providing apparatus that provides information related to content that a user is interested in to the terminal device of the user, and holds a plurality of types of content description information that describes a description of a plurality of types of content. Content description information management means, user interest information management means for holding user interest information that is information related to the interests of a plurality of users, and a plurality of types of content description information held by the content description information management means, The content description information is selected by referring to the content description information selection means for selecting and reading the content description information used to determine the content that the user is interested in, and the user interest information held in the user interest information management means. Based on the content description information read by the selection means Te, characterized in that it comprises a providing means for providing the recommended content determining means for determining a content to be recommended to the user, information relating to the determined recommended content to the terminal device of the user. As a result, the recommended content is determined from the content description information selected from the plurality of content description information and the user interest information, so that the recommended content accurately reflecting the user's interest is provided.

  More specifically, the information providing apparatus according to the present invention is an information providing apparatus that provides information related to content that a user is interested in to the terminal device of the user via a communication network, and a plurality of different types of terminals. Content interest information management means for holding a plurality of pieces of content explanation information describing descriptions about a plurality of pieces of content associated with each device, and user interest information management for holding user interest information which is information relating to the interests of a plurality of users A plurality of pieces of content description information stored by the content description information management unit based on a type of the terminal device of the accessing user and a receiving unit that accepts access by the user terminal device via the communication network. From the recommendations used to determine the content that the user is interested in Access is made by referring to content description information selection means for selecting and reading content description information and providing content description information for provision to the user, and user interest information held in the user interest information management means. Recommended content determining means for determining information to be recommended to the user based on the identified information and the recommended content description information read by the content description information selecting means; Providing means for specifying information relating to the determined recommended content from the providing content explanation information read by the content explanation information selecting means, and providing the identified information to the terminal device of the user. And As a result, the content description information used to determine the content recommended for the user is separate information different from the content description information provided to the user's terminal device, so the type of the terminal device that has accessed the information providing device The content to be recommended to the user is appropriately determined without depending on. Since the content explanation information provided to the user is used in correspondence with the type of the terminal device of the user, for example, compact information or detailed information is selected depending on whether it is a mobile phone or a PC. Provided.

  For example, TV programs broadcast in many countries and languages have abundant information on the language of the country in which they were produced, such as “title”, “performers”, and “program details”, and information in other languages is “titles” There is only that. In this case, content description information in a language from which abundant description information can be obtained is used as content description information used for learning and recommendation, and content description information using the user's native language is used for display to the user. Thus, even when there is not sufficient content description information in the user's native language, it is possible to make a recommendation with higher accuracy.

  Here, the content description information selection means has, for example, a selection database that is a table in which the types of terminal devices are associated with recommendation content description information and provision content description information, and refers to the selection database. The recommendation content description information and the provision content description information may be selected according to the above, and the selection database may include the type of terminal device, the recommendation content description information, and the provision content description for each of a plurality of users. The content description information selecting means refers to the selection database, and the recommended content description information corresponding to a combination of the accessing user and the type of the terminal device of the user, The providing content description information may be selected. As a result, not only the type of the terminal device but also the distinction of the user is considered, and the recommended content explanation information and the provision content explanation information are selected. Therefore, the content explanation information more accurately reflects the user's preference and interest. Provision is possible.

  As a method for determining recommended content, for example, the user interest information includes an interest keyword indicating a term of interest for each of the plurality of users, and the recommended content description information includes the plurality of content items. For each content, an explanation keyword indicating a term related to the explanation of the content is included, and the recommended content determination means determines the content by matching the interest keyword corresponding to the accessing user with the explanation keyword. May be.

  The information providing apparatus further includes user history information input means for acquiring history information related to content selection by the user, and user interest information held in the user interest information management means based on the acquired history information. Learning means for updating. At this time, the content description information selection means further updates the user interest information from among the plurality of content description information stored by the content description information management means based on the type of the terminal device of the accessing user. The learning content description information used for the selection is read out, and the learning means may update the user interest information based on the content description information read by the content description information selection means and the history information. Good. Thereby, the content description information for learning the user interest information is selected and used separately from the recommendation content description information and the provision content description information, so that the accuracy of the user interest information is surely ensured as it is used. It will be improved.

  As a learning method, for example, the user interest information includes an interest keyword indicating a term of interest for each of the plurality of users and weight information regarding the weight, and the history information includes content selection. Weight information indicating the degree of interest of the user is included, the recommended content determination means determines the content based on an interest keyword and weight information corresponding to the accessing user, and the learning means includes the history The weight information of the interest information may be updated based on the weight information included in the information.

  The content description information management means includes a content description information creation unit that creates new content description information from the plurality of content description information, and the content description information selection means is created by the content description information creation unit. The recommendation content description information and the provision content description information may be selected from a plurality of content description information including the content description information. Specifically, the content description information creating unit may create the content description information by taking the sum or product of information included in the plurality of content description information. As a result, new content description information is generated from the already existing content description information, and it is possible to provide content description information reflecting a variety of terminal devices and user environments and preferences.

  In addition, the content description information selection unit acquires switching content description information used for switching the content description information from among the plurality of content description information held in the content description information management unit. And a switching determination unit that determines whether to switch the recommended content description information, the providing content description information, and the learning content description information to be selected based on terms included in the acquired switching content description information. And a post-switching content description information acquisition unit that selects and acquires new recommendation content description information, provision content description information, and learning content description information when it is determined to switch. Specifically, the switching content description information includes a keyword indicating a term describing the content, and the content description information selecting means further includes a switching keyword and a recommended content after switching for a plurality of users. A switching selection database indicating correspondence between the description information, the providing content description information, and the learning content description information is held, and the switching determination unit is included in the keyword included in the switching content description information and the switching selection database. The above determination may be made by collating with the switching keyword. Thereby, according to the user's preference and the like, it is possible to switch the recommended content explanation information, the provision content explanation information, and the learning content explanation information, so that the content explanation information more reflecting the user's preference and interest can be provided. It becomes possible.

  The recommended content determining means determines the content according to any of a plurality of predetermined recommendation methods, and the content description information selecting means is a plurality of content descriptions held in the content description information managing means. Among the information, the switching content explanation information acquisition unit for acquiring the switching content explanation information used for switching the recommendation method used for determining the recommended content by the recommended content determination means, and the acquired switching content A switching determination unit that determines whether or not to switch the recommendation method based on a term included in the description information, and the recommended content determination unit performs switching when the switching determination unit determines to switch The content is determined according to a new recommendation method later, The means updates the user interest information according to any of a plurality of predetermined recommendation methods, and the content description information selection means is selected from the plurality of content description information held in the content description information management means. , Included in the acquired switching content description information, and a switching content description information acquisition unit that acquires switching content description information used for switching the recommendation method used to update the user interest information by the learning means. A switching determination unit that determines whether or not to switch the recommendation method based on a term, and when the learning unit determines that the switching determination unit switches, according to the new recommendation method after switching The user interest information may be updated. As a result, it is possible to switch between an evaluation method for determining recommendation content and an evaluation method for learning and updating user interest information according to the user's preference and the like. It is possible to provide reflected content description information.

  In the information providing apparatus, the content is a program, the content description information is program information describing a description about the program, and the recommended content determination unit selects a recommended program as content recommended to the user. The providing means may determine and provide information on the recommended recommended program. In this case, the information providing apparatus further includes a recommended program operation history storage unit that stores the user operation history for the recommended program determined by the recommended content determination unit, and the content explanation information selection unit includes the recommended program. It is preferable to select program information from the operation history of the user stored in the operation history storage means. Accordingly, when there is a plurality of program information, even if there is a difference in the programs recommended by the respective program information, the program information that can make an optimum program recommendation by using the user operation history for the recommended programs. Can be selected automatically.

  In addition, the information providing apparatus further includes a recommended program operation history storage unit that stores the user's operation history for the recommended program determined by the recommended content determination unit, and the recommended program operation history storage unit that stores the recommended program operation history storage unit. A recommended condition determining means for determining a recommended condition for a program based on a user operation history, wherein the recommended content determining means determines a recommended program that satisfies the recommended condition determined by the recommended condition determining means. Good. As a result, when there is a plurality of program information, even if there is a difference in the programs recommended by the respective program information, for example, the recommended program is preferentially assigned to each program information using the user's recording history. It is possible to automatically determine the conditions to be performed and realize highly accurate program recommendation.

  The information providing apparatus further includes program information generating means for generating a new program information by combining a plurality of types of program information stored in the content description information managing means, and the content description information selecting means comprises: The program information used by the recommended content determination unit may be selected from a plurality of types of program information stored in the content description information management unit and the program information synthesized by the program information synthesis unit. As a result, when there is a plurality of program information, it is possible to synthesize the program information to generate new program information and to recommend as wide a range of program information as possible to the user. Furthermore, by using the history that the user has viewed and recorded from the recommended programs, it becomes possible to automatically select program information when providing recommended programs to the user.

  The information providing apparatus further selects a user operation history storage unit that stores the user operation history for the program, and a specific type of operation history from the operation history stored in the user operation history storage unit, A configuration comprising content information learning means for learning user preferences based on the selected operation history, wherein the user interest information management means retains user interest information indicating user preferences learned by the content information learning means It is good. As a result, in the initial stage of learning, it is possible to recommend programs using program information that has been operated by the user and to recommend as much program information as possible. Furthermore, it is possible to automatically select an operation history to be used for recommending a program by using a user's browsing / operation history of the recommended program.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(Embodiment 1)
FIG. 1 is a block diagram showing a configuration example of an information providing system 1 according to Embodiment 1 of the present invention. This information providing system 1 is a system in which recommended content information is the same regardless of a terminal operated by a user and a terminal browsed, and includes a content explanation information management device 11, a user interest information management device 12, and a content recommendation device. 13 and a terminal 14.

  The content explanation information management device 11 is a computer device that stores a plurality of pieces of content explanation information, that is, content explanation information, for example, an Internet information website. The content explanation information in the present embodiment is a database that stores information that describes the content that is the target in the information providing system 1. For example, if the content is a television program, an electronic program guide using the Internet network is used. (EPG), an electronic program guide using VBI of television broadcasting, an impression of a television program created by an individual, and the like.

  Here, when the terminal targeted by the content description information 111 is a PC (personal computer), as shown in FIG. 2, the content ID 111a for identifying the content, the content name 111b for each content ID 111a, and the content are broadcast. Information such as the channel 111c to be played, the broadcast start date 111d of the content, the temporary broadcast 111e at the end of the broadcast of the content, the content description document 111f such as the content and impression of the content, and the content description keyword 111g which is a keyword characterizing the content. Yes. Note that information may not necessarily be registered for items other than the content ID 111a, the content description document 111f, and the content description keyword 111g.

  Similarly, when the terminal targeted by the content description information 112 is a mobile phone connected to the Internet, as shown in FIG. 3, the content ID 112a for identifying the content, the content name 112b for each content ID 112a, the content Information such as the channel 112c on which the content is broadcast, the content broadcast start date and time 112d, the content broadcast temporarily 112e at the end of the content broadcast, the content description document 112f such as the content and impression of the content, and the content description keyword 112g that is a keyword characterizing the content are stored. is doing. As with the content description information 111, information other than the content ID 112a, the content description document 112f, and the content description keyword 112g may not necessarily be registered. In addition, a content description document may not be registered in content description information that is used only for recommendation and learning. In addition, when content description information does not contain content description information, other content description information may be used.

  The user interest information management device 12 can be accessed from interest information indicating what kind of program the user is interested in, that is, a computer device storing the user interest information 121, for example, a storage on the Internet or the outside. Home server in a household.

  The user interest information 121 is a database that stores user interest information of a target user in the information providing system 1. As shown in FIG. 4, a user ID 121a for identifying a user, a keyword or genre in which the user is interested or not interested, a keyword 121b indicating a performer, and automatically calculated from user history information, or Information such as a keyword weight 121c indicating an evaluation value for each keyword 121b, which is manually input, is stored. For example, a user whose user ID is “100” is interested in a program with keywords “future”, “universe”, “drama”, and “pro-wrestling”, and in particular, “pro-wrestling” and “future” are important. Show. Note that information may not necessarily be registered for items other than the user ID 121a and the keyword 121b.

  The content recommendation device 13 learns user interest information from the content explanation information and the user history using a statistical method, and provides a recommended content from the content explanation information and the user interest information, for example, on the Internet This is a website, and includes a content information selection unit 131, a content information recommendation unit 132, a recommended content provision unit 133, a content information learning unit 134, a user history information input unit 135, and a selection database 1310.

  The content information selection unit 131 selects content description information to be used based on the selection database 1310 when providing content, recommending content, or learning user interest. As shown in FIG. 5, the selection database 1310 is used when recommending content information from a terminal ID 1311 a for identifying a terminal, providing content explanation information 1311 b used when providing recommended content information, and user interest information 121. Information such as recommended content explanation information 1311c and learning content explanation information 1311d used when learning user interest information 121 is stored. For example, when the terminal ID is “mobile phone 2”, “mobile phone EPG” is selected as the providing content description information, “PC EPG” is selected as the recommended content description information, and learning content The description information indicates that “TV EPG” is selected.

  The content information recommendation unit 132 calculates a content evaluation value from the recommendation content description information 1311 c selected by the content information selection unit 131 and the user interest information 121.

  The recommended content providing unit 133 determines recommended content based on the content evaluation value calculated by the content information recommendation unit 132, and provides the providing content description information 1311b to the user.

  The user history information input unit 135 receives user histories such as user evaluations for recommended content, operation histories such as content viewing and recording, and the like from the terminal.

  Here, as shown in FIG. 6, the user history information input from the terminal includes a user ID 117a for identifying the user, a content ID 117b for identifying the content, user's evaluation of the content, viewing of the content, and operations such as recording. Information such as a user history 117c such as a history, user evaluation, user history date 117d indicating the date of operation history, and the like is stored. Note that information may not necessarily be input for items other than the user ID 117a, the content ID 117b, and the user history 117c.

  The content information learning unit 134 learns user interest information from the learning content explanation information selected by the content information selection unit 131 and the user history obtained by the user history information input unit 135.

  The terminal 14 is connected to a network such as a TV, a PC, a mobile phone, or a PDA (Personal Digital Assistance) such as a TV or PC, and a terminal that provides recommended content information from the content recommendation device 13. It is. Here, 141 is a mobile phone, 142 is a TV, and 143 is a PC.

Next, the operation of the information providing system 1 configured as described above will be described.
FIG. 7 is a flowchart showing an operation procedure of the information providing system 1 until the recommended content information is presented to the user. First, a user sends a terminal ID and a user ID from the terminal 14 to the content recommendation device 13 via the network, and the content recommendation device 13 uses a terminal ID and a user ID at a terminal and a user ID input unit (not shown). Is received (S701). For example, when the terminal 14 is a PC and accesses the content recommendation device 13 using a browser, the content recommendation device 13 transmits a login screen to the user terminal in response to an access request from the terminal 14. The terminal 14 that has received the login screen transmits the terminal ID and the user ID to the content recommendation device 13 in response to a terminal ID and user ID transmission request from the user. When a browser is used, HTTP_Agent indicating the terminal type may be used as the terminal ID. Further, the user ID may be stored in the cookie, and the contents of the cookie may be used without inputting the user ID.

  Subsequently, the content information selection unit 131 determines recommended content description information from the selection database 1310 using the terminal ID as a search key (S702). For example, in FIG. 5, when the terminal ID is “TV2”, “PC EPG” is selected as the recommended content explanation information 1311c. Subsequently, the content information selection unit 131 determines providing content description information from the selection database using the terminal ID as a search key (S703). For example, in FIG. 5, when the terminal ID is “mobile phone 1”, “mobile phone EPG” is selected as the providing content description information 1311b.

  The content information recommendation unit 132 reads the user interest information 121 from the user interest information management apparatus 12 using the user ID as a search key. For example, in FIG. 4, when the user ID is “UID100”, the keywords 121 b and the keywords “future, 10.5”, “universe, 7.2”, “drama, 5.6”, “pro-wrestling, 12.3” A combination of weights 121c is read. Subsequently, the evaluation value of each content is calculated from the recommended content explanation information selected by the content information selection unit 131 and the user's interest information (S704), and the recommended content is determined from the evaluation value. For example, when the content description information shown in FIG. 2 is recommendation content description information and the user ID is UID100, the evaluation value with the content ID V010M001 is calculated as the content of “future”, “story”, and “universe”. The explanation keyword 111g is compared with the user interest information, and “17.7” is obtained from the sum of the keyword weights of “future” and “universe” included in the user interest information. Evaluation values are similarly calculated for other content IDs, and the top 10 evaluation values are set as recommended contents. In addition, since there is a possibility that the content with the higher evaluation value is only similar content, the recommended content is a total of 10 items with the upper evaluation value and the middle evaluation value of 5 cases. Only the content with the higher evaluation value may not be used.

  Further, the number of recommended programs may be changed according to the display performance of the terminal. For example, by providing display terminal performance storage means for storing the performance of the display terminal, if the display terminal has insufficient resolution, the number of recommended programs is reduced, and the display terminal having a large screen such as a stationary television In this case, the number of recommended programs can be increased. As a result, the programs recommended to the upper level in each terminal are common, and an appropriate number of programs can be recommended according to the terminals.

  Also, by providing recording capacity detecting means for detecting the recording capacity from the capacity of the storage medium and the capacity of the already recorded program, the recommended number of programs is determined according to the capacity of the storage medium such as a hard disk for recording the program. You can also As a result, when all the recommended programs are to be recorded, it is not possible to record due to a lack of recording medium capacity.

  In addition, by providing a recording history storage means for storing the recording history of the user, the history of the recording reservation by the user is stored, and by detecting how far the recording reservation is made up to the previous program. The recommended program can be determined according to the number of recommended programs and the broadcast date. For example, when a user usually does not reserve a program for one week ahead, the user decides a recommended program from among the programs for one week ahead. On the other hand, it is also possible to determine a program to be recommended from programs up to a few days ahead of the program recommendation date for a user who sets a reserved recording only for a program a few days ahead. This makes it possible to perform program recommendation according to the program recording style of each user. Furthermore, when a recommended program is determined according to each terminal, the recommended program period may be determined from the history of reserved recording at each terminal. For example, while a mobile phone only schedules recording for programs up to a few days in advance, a home video recording program for a user who schedules recording up to a week in advance for each user has a scheduled recording history at each terminal. The period of the recommended program may be determined according to the above. In addition, if scheduled recording of a program becomes frequent during the year-end and New Year holidays or program modification time, it is possible to recommend a program earlier than usual.

  In the initial stage when the user starts using the information providing system, if the user's history is not sufficient and the user's interest information has not been extracted, the program has a lot of content explanation information, and the audience rating is high. If a general content such as a program or a program for which many other users make recording reservations is recommended and the history is sufficiently accumulated, the recommended content may be determined from the interest information of the user. At this time, a general program with a high audience rating and a program recommended from user's interest information are displayed well in the world by distinguishing them by changing colors or changing the font size. It is possible to distinguish and check a program that is currently being viewed and a program that is of high interest to the user. Thus, although the initial stage is only general programs, the recommendation of programs according to the user's interest information increases as the user's history is accumulated.

  Then, the recommended content providing unit 133 provides the content description information for provision corresponding to the recommended content determined by the content information recommendation unit 132 to the user's terminal (S705). As a result, for example, when the user ID is “UID100”, the recommended content description information is the first to sixth lines in FIG. 2, and the recommended content displays the top three evaluation values, the content The content IDs recommended by the information recommendation unit 132 are “V010M001”, “V010M002”, and “V411B000”. When the content explanation information for provision is “EPG for mobile phone” shown in FIG. 3, from this content ID, “Future Story 1—Far Way” — “Future Story 2—Follow Way”, “Laughter Pro Wrestling—Laughter Is provided to the terminal.

  The determination of the content to be provided (S703) may be performed anywhere after obtaining the terminal ID and the user ID from the terminal (S701) and before providing the recommended content to the terminal (S705).

  FIG. 8 is a diagram illustrating a communication sequence in the information providing system 1 until the recommended content information is presented to the user. The user transmits a terminal ID and a user ID from the terminal 14 to the content recommendation device 13 via the network (S801). Next, the content information selection unit 131 determines recommended content description information from the selection database 1310 using the terminal ID as a search key (S802), and requests the content description information management apparatus 11 for content description information. Subsequently, the content information selecting unit 131 determines providing content description information from the selection database using the terminal ID as a search key (S803), and requests the content description information management apparatus 11 for the content description information. Next, the content information recommendation unit 132 requests the interest information of the corresponding user from the user ID to the user interest information management device.

  The content information recommendation unit 132 calculates the evaluation value of each content from the recommendation content explanation information and the user's interest information (S804), and determines the recommended content from the evaluation value.

  Then, the recommended content providing unit 133 provides providing content description information corresponding to the recommended content determined by the content information recommendation unit 132 to the user terminal (S805).

  FIG. 9 is a flowchart showing an operation procedure of the information providing system 1 until learning of user interest information from the content explanation information and the user history. First, a user sends a terminal ID, a user ID, and a user history from the terminal 14 to the content recommendation device 13 via the network, and the content recommendation device 13 uses a terminal and a user ID input unit (not shown). The terminal ID and the user ID are received, and the user history is received by the user history information input unit 135 (S901). For example, when the terminal 14 is an HDD recorder and accesses the content recommendation device 13, the terminal ID, the user ID, and the user's history information such as a recording operation and an input of evaluation are transmitted from the terminal 14 to the content recommendation device 13. Send.

  Subsequently, the content information selection unit 131 determines learning content explanation information from the selection database 1310 using the terminal ID as a search key (S902). For example, in FIG. 5, when the terminal ID is “HDD2”, “TV EPG” is selected as the learning content explanation information 1311d.

  The content information learning unit 134 calculates a keyword evaluation value from the learning content explanation information and the user history information (S903). Then, the content information learning unit 134 requests the user interest information management apparatus 12 to update the user interest information obtained from the user history input (S904). For example, in the keyword weight learning, it is assumed that “video recording” has a keyword weight of +0.5, and “interested” has a keyword weight of +1.0. Here, the content explanation information for learning is FIG. 2, and when there is an input of the user's history as in the first line of FIG. 6 and there is an input of “recording” + “interested”, the content ID For “V145D001”, a combination of a keyword 121b and a keyword weight 121c of “drama, +1.5”, “story, +1.5”, “royal road, +1.5” and user interest information obtained from the user's history input Become. In the update of the user interest information 121, when the keyword “drama” already exists in the user interest information 121, the keyword weight of the keyword “drama” is set to “+1.5”, and the keyword “royal road” is the user interest information. If it does not exist in 121, a combination of the keyword “royal road” and the keyword weight “1.5” is added.

  Moreover, although the evaluation value of the keyword was calculated from the user history and the content explanation information, the user interest information may also be learned using the user interest information before learning.

  FIG. 10 is a diagram showing a communication sequence in the information providing system 1 until learning user interest information from content explanation information and a user history. First, a user transmits a terminal ID, a user ID, and a user history from the terminal 14 to the content recommendation device 13 via the network (S1001). Next, the content information selection unit 131 determines learning content description information from the selection database 1310 using the terminal ID as a search key (S1002), and requests the content description information from the content description information management apparatus 11. Next, the content information learning unit 134 calculates a keyword evaluation value from the learning content explanation information and the user history information (S1003), and from the keyword evaluation value obtained from the user ID and the user history input, A request for updating the user interest information 121 of the user interest information management apparatus 12 is made (S1004).

  FIG. 11 is a diagram for explaining the operation when the user interest information management device 12 and the content recommendation device 13 are arranged on a home network. Here, user histories are collected and accumulated as user interest information, EPGs for PC / TV / mobile phone are rearranged by interest information, and accumulated as EPG for recommendation & learning, and the user interest information In addition, it is shown that a recommended program is displayed on the user's terminal based on the EPG for recommendation & learning. In this way, by configuring the information providing system in the home, it is possible to obtain recommended content information without notifying the content explanation information provider of user interest information that is personal information.

  FIG. 12 is a diagram for explaining the operation when the providing device of the content explanation information provider intervenes between the terminal and the content recommendation device. Here, a state in which a portable program information providing site, a TV program information providing site, and a PC program information providing site are interposed between the terminal and the content recommendation device is shown. In this case, if the content explanation information provider sends the user's history to the content recommendation device, and the content recommendation device sends the recommended content information (only the recommended content ID may be used) to the content explanation information provider, the user However, even if a plurality of content explanation information providers are used, the same recommended content information can be obtained.

  Thus, according to the information providing system 1 in the present embodiment, the user can obtain the same recommended content information regardless of the operated terminal and the browsed terminal. That is, regardless of the content description information to be displayed, the same content is recommended regardless of whether the user's terminal is a mobile phone or a PC. Further, even if the user inputs content evaluation with the PDA and views the content with the TV, the history information of the same user is input from the content recommendation device, and there is a bias in learning by the terminal. It does not occur and user history information can be managed in a unified manner.

  Further, although the selection database 1310 has the content information for recommendation and the content description information for learning different, as shown in FIG. 13, the content information for recommendation and the content description information for learning use the same information. Also good. FIG. 13 shows an example of a selection database in which the content description information for recommendation and learning is indicated by the common item 1311f.

Further, the selection database 1310 can select content explanation information for provision, recommendation, and learning by a terminal, not by a user. As shown in FIG. 14, the user can select content explanation information by a terminal. May be selected. FIG. 14 shows an example of a selection database provided with a user ID 1311e. By enabling the user to select the content description information, the user's favorite content description information can be used regardless of the type of terminal. Also in this case, as shown in FIG. 15, the recommendation content description information and the learning content description information may be used in the same manner. FIG. 15 shows an example of a selection database provided with a user ID 1311e and recommendation and learning content explanation information 1311f. In this embodiment, the content is a TV program, but music, movies, paintings, etc. The same effect can be obtained if a plurality of explanations and comments can be obtained for one content.

  In addition, by replacing content with news sources such as incidents and events, it becomes possible to replace content explanation information with newspaper articles, reviews of events, commentary, etc., and the present invention is not only for content but also for information in general. The same effect can be obtained.

  Further, the recommendation method is not limited to the contents described in the present embodiment, and any recommendation method may be used as long as it is a content-based filtering method that uses content description information.

  In addition, when the content description information provided by the content description information provider does not include a content description keyword, the keyword may be extracted and generated from the content description document by using morphological analysis. Further, the frequency in the keyword document may be registered in the content explanation information and used for recommendation and learning.

  In addition, when content description information is provided from a plurality of content description information providers, content description information having the most detailed content description information for one content regardless of the content description information provider May be used.

  Similarly, content description information may be newly created from each content description information. For example, content description information may be newly created and used based on the sum of the content description information (logical sum of content description keywords). 16 creates content description information 113 (information composed of items 113a to 113g similar to the content description information 111 and 112) based on the sum of the content description information 111 of FIG. 2 and the content description information 112 of FIG. In this example, the frequency of content explanation keywords is also registered.

  Similarly, content description information may be newly created from content description information included in all content description information and used (logical product of content description keywords). 17 creates content description information 113 (information composed of items 113a to 113g similar to the content description information 111 and 112) by the product of the content description information 111 of FIG. 2 and the content description information 112 of FIG. It is an example.

  In addition, when content description information is provided from a plurality of content description information providers, the user can select content description information, so that user-specific content description information can be selected from the user's selection. You may make it produce (creation of content description information from a user history). Also, the content description information of the broadcast content differs in the content description contents one month before, one week before, and one day before the program is distributed. Therefore, it is possible to process these pieces of explanation information as a plurality of pieces of content explanation information for one content. For example, when the content explanation information is more detailed one day before than one month ago, it is possible to accurately reflect the user's interest by using the content explanation information one day ago for learning. Become.

  In this embodiment, the content ID is managed in a unified manner among a plurality of content explanation information providers. However, the content ID may not be managed in a unified manner. For example, in the case of a television program, the content ID can be managed by handling the title ID in a unified manner because the channel and the broadcast time coincide. However, if the channels match, but the broadcast start time and the broadcast end time do not match completely, the broadcast start time must match, 90% of the broadcast time must match, It is possible to handle content IDs in a unified manner by using the degree of similarity between content description information as to whether content matches, such as 70% match. In addition, one program has one piece of content explanation information. However, when a plurality of programs are explained by one piece of content explanation information, another program has one piece of content explanation information. The content description information of the content description information provider may be used.

  In this embodiment, the program is recommended only from the content explanation information. However, the program may be recommended in consideration of time. For example, for programs with different broadcast times but similar contents and the same content explanation keywords, programs with close broadcast times may be recommended to the top.

  Further, even if the content explanation information of the program is up to two months ahead, if the recording apparatus can make a recording reservation only up to one month ahead, the recommendation may be made from the content explanation information for one month.

(Embodiment 2)
Next, a second embodiment of the present invention will be described. FIG. 18 is a block diagram showing a configuration example of the content recommendation device 13a of the information providing system 2 according to Embodiment 2 of the present invention. The information providing system 2 has basically the same system configuration as that of the first embodiment, and includes a content recommendation device 13a, a content explanation information management device 11, a user interest information management device 12, and a communication network 10. The terminal 14 is connected to the content recommendation device 13a via the terminal, but in this figure, the user interest information management device 12, the communication network 10, and the terminal 14 are not shown.

  The content recommendation device 13a is the content recommendation device 13a in the information providing system in which the recommended content information is the same regardless of the terminal operated by the user or the terminal browsed. The content recommendation device 13a, the content information selection unit 131a, the content information recommendation unit 132, A recommended content providing unit 133, a content information learning unit 134, a user history information input unit 135, and a selection database 1310 are included.

  This content recommendation device 13a basically has the same configuration as the content recommendation device 13 in the first embodiment. However, the content information selection unit 131a is different from the first embodiment in that the content information selection unit 131a includes a switching KW content explanation information acquisition unit 1311, an information switching KW determination unit 1312, and a switching selection database 1316. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  The content information selection unit 131a is a unit that selects content description information to be used for each user by using a keyword included in the content description information when providing the content, recommending the content, or learning the user's interest. The switching KW content description information acquisition unit 1311, the information switching KW determination unit 1312, the recommendation content description information acquisition unit 1313, the learning content description information acquisition unit 1314, and the provision content description information acquisition unit 1315. And a switching selection database 1316.

  The switching KW content description information acquisition unit 1311 acquires content description information from the content description information management apparatus 11 and sends a content description keyword of the acquired content description information to the information switching KW determination unit 1312.

  The information switching KW determination unit 1312 provides content explanation information to be used for providing content, recommending content, learning user's interest, user ID, and switching KW content explanation information acquisition unit 1311. Is selected based on the content explanation keyword obtained from the above and the switching selection database 1316.

  As shown in FIG. 19, the switching selection database 1316 includes a user ID 1316 a for identifying a user and a content explanation information keyword acquired by the content explanation information acquisition unit 1311 for switching KW. , Switching KW 1316b for switching content explanation information for provision, recommendation content explanation information 1316c used when recommending content information from user interest information 121, and learning content explanation information 1316d used when learning user interest information 121 In addition, information such as providing content explanation information 1316e used when providing recommended content information is stored.

  The recommended content description information acquisition unit 1313 acquires the recommended content description information selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

  The learning content description information acquisition unit 1314 acquires the content description information for learning selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

  The providing content description information acquisition unit 1315 acquires the providing content description information selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

Next, the operation of the information providing system 2 configured as described above will be described.
FIG. 20 is a flowchart showing an operation procedure of the information providing system 2 until the recommended content information is presented to the user. The operation of the information providing system 2 is basically the same as the operation until the recommended content of the information providing system 1 in Embodiment 1 is provided to the user. However, the step of determining recommended content description information (S702) and the step of determining providing content description information (S703) are different from those of the first embodiment. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  First, the user sends a terminal ID and a user ID from the terminal 14 to the content recommendation device 13a via the network, and the content recommendation device 13a receives the terminal ID and the user ID at the terminal and the user ID input unit (S701).

  Subsequently, the content description information is acquired by the switching KW content description information acquisition unit 1311 of the content information selection unit 131a. The content description information at this time may be determined by the content description information provider or may be determined for each user. The content description keyword of the acquired content description information is sent to the information switching KW determination unit 1312. The information switching KW determination unit 1312 determines whether the user ID and the content explanation keyword are registered in the switching selection database 1316 using the search key as a search key (S702a). When the content description keyword includes the switching KW, the recommendation content description information registered in the switching recommendation content description information in the switching selection database 1316 is selected (S702b). For example, in FIG. 19, when the user ID is “UID 523” and the content explanation keyword includes “Western picture”, “PC EPG” is selected as the recommended content explanation information 1316c regardless of the selection database 1310. Is done. Next, the recommendation content description information acquisition unit 1313 acquires the recommendation content description information selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

  Subsequently, the providing content description information registered in the switching providing content description information is also selected (S703b). For example, in FIG. 19, when the user ID is “UID 523” and the content explanation keyword includes “animation”, “EPG for mobile phone” is selected as the provision content explanation information 1316d regardless of the selection database 1310. Is done. Next, the providing content description information acquisition unit 1315 acquires the providing content description information selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

  Similarly, when the content description keyword does not include the switching KW, the recommendation content description information registered in the selection database 1310 is selected (S702c), and then the provision content description information is selected (S702c). S703c).

  The content information recommendation unit 132 reads the user interest information 121 from the user interest information management apparatus 12 using the user ID as a search key. Subsequently, an evaluation value of each content is calculated from the recommended content description information selected by the content information selection unit 131a and the user's interest information (S704), and the recommended content is determined from the evaluation value.

  Then, the recommended content providing unit 133 provides the content description information for provision corresponding to the recommended content determined by the content information recommendation unit 132 to the user's terminal (S705).

  FIG. 21 is a flowchart showing an operation procedure of the information providing system 2 from learning of user interest information from content explanation information and user history. The operation of the information providing system 2 is basically the same as the operation of learning the user interest information from the content explanation information and the user history of the information providing system 1 in the first embodiment. However, the step of determining learning content explanation information (S902) is different from the first embodiment. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  First, a user sends a terminal ID, a user ID, and a user history from the terminal 14 to the content recommendation device 13a via the network, and the content recommendation device 13a uses a terminal and a user ID input unit (not shown). The terminal ID and the user ID are received, and the user history is received by the user history information input unit 135 (S901).

  Subsequently, the content description information is acquired by the switching KW content description information acquisition unit 1311 of the content information selection unit 131a. The content description information at this time may be determined by the content description information provider or may be determined for each user. The content description keyword of the acquired content description information is sent to the information switching KW determination unit 1312. The information switching KW determination unit 1312 determines whether or not the user ID and the content explanation keyword are registered in the switching selection database 1316 using the search key (S902a). If the content description keyword includes the switching KW, the learning content description information registered in the switching learning content description information in the switching selection database 1316 is selected (S902b). For example, in FIG. 19, when the user ID is “UID444” and the content description keyword includes “baseball”, “TV EPG” is selected as the learning content description information 1316d regardless of the selection database 1310. Is done. Similarly, if the content description keyword does not include the switching KW, the learning content description information registered in the selection database 1310 is selected (S902c). Next, the learning content description information acquisition unit 1314 acquires the learning content description information selected by the information switching KW determination unit 1312 from the content description information management apparatus 11.

  The content information learning unit 134 calculates a keyword evaluation value from the learning content explanation information and the user history information (S903). Then, the content information learning unit 134 requests the user interest information management apparatus 12 to update the user interest information obtained from the user history input (S904).

  As described above, according to the information providing system 2 in the present embodiment, the user can obtain the same recommended content information regardless of the operated terminal and the browsed terminal. If there is a difference, the content description information specified by the user can be used regardless of the terminal. In other words, regardless of the content description information, regardless of the content description information, the same content is recommended, and even if the user's terminal is a terminal with a small display area such as a mobile phone, the content that the user is interested in More detailed description information can be provided. Furthermore, the description of the content can be changed according to the user's preference.

  In the present embodiment, the content description information is switched according to the keyword in which the user is interested. Furthermore, content description information suitable for each user may be selected from the user's operation history for the program recommended to the user. The information stored in the content description information 111, 112, 113 in FIG. 22 is different description information for the same program content. Therefore, three types of keywords can be extracted from the same operation history for the user's program. For example, as shown in FIG. 22, there are cases where the keywords estimated to be of interest to the user are different because the content description information is different from the operation history for the same program of the user whose user ID is UID 523. The user selects a program with interest in a baseball program, especially for a program in Kansaka. At that time, general explanation information is given to the program in the content explanation information 111, specific explanation information is given in 112 by kanji, etc., and in 113, the explanation information is given by alphabet, katakana, etc. Has been granted. Depending on these extracted keywords, the programs that the user will be interested in are also different. Therefore, as shown in FIG. 22, there are recommended programs and the same programs, but there are also different programs.

  Therefore, all the programs extracted in FIG. 22 are presented to the user. However, the user performs scheduled recording of the programs “BaseBall Magazine” and “Tigers Information” with respect to the presented program, and does not view “Journey of Kansaka Train”. That is, since the user's interest has been confirmed for the program recommended using the content description information 113, it is determined that it is appropriate to use the content description information 113 for this user. On the other hand, it is determined that an erroneous recommendation has been made for the program “Trip of the Kansaka Train” recommended by the content description information 112, and it is determined that the content description information 112 is not suitable for this user. be able to.

  Thus, when there are a plurality of program explanation information for one program, it is determined which content explanation information is suitable from the user's operation history with respect to the program recommendation result, and the optimum content. Explanation information can be selected.

  The switching selection database 1316 may be set by the user himself / herself or may be set by the content explanation information provider.

  Further, the switching selection database 1316 has the recommendation content description information and the learning content description information different from each other, but the same information may be used for the recommendation content description and the learning content description information.

  Further, although the switching KW 1316b of the switching selection database 1316 is a keyword, it may be a keyword logical expression expressed by logical sum or logical product, and is not limited to a keyword such as genre information possessed by content.

  Further, although the information switching KW determination unit 1312 determines whether or not there is a keyword, when the degree of interest in the user's content, that is, the evaluation value of the content is used and the value is equal to or greater than a certain value, You may make it switch provision content description information. Similarly, the provided content description information may be switched when it has a value less than or equal to a certain value.

(Embodiment 3)
Next, a third embodiment of the present invention will be described. FIG. 23 is a block diagram illustrating a configuration example of the content recommendation device 13b of the information providing system 3 according to Embodiment 3 of the present invention. The information providing system 3 includes a content recommendation device 13b, a content explanation information management device 11, a user interest information management device 12, and a terminal 14 connected to the content recommendation device 13b via the communication network 10. However, in this figure, illustration of the content description information management device 11, the user interest information management device 12, the communication network 10, and the terminal 14 is omitted.

  The content recommendation device 13b is a content recommendation device 13b in the information providing system in which the recommended content information is the same regardless of the terminal operated by the user and the terminal browsed. The content recommendation device 13b includes a content information selection unit 131b and a plurality of content information recommendation units. 132a and 132b, a recommended content providing unit 133, a plurality of content information learning units 134a and 134b, a user history information input unit 135, and a selection database 1310.

  The content recommendation device 13b basically has the same configuration as the content recommendation device 13 in the first embodiment. However, the content information selection unit 131b includes a switching KW content explanation information acquisition unit 1311, a recommendation switching KW determination unit 1312a, and a recommendation method selection database 1317, and the content information recommendation unit 132 and the content information learning unit in the first embodiment. Instead of 134, the embodiment is different from the first embodiment in that it includes a combination of not only one but a plurality of content information recommendation units 132a and 132b and content information learning units 134a and 134b. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  The content information selection unit 131b is a unit that selects a recommendation method to be used when recommending content by using a keyword included in the content description information for each user. The content information selection unit 131b and the recommended switching KW determination A section 1312a, a recommendation content description information acquisition section 1313, a learning content description information acquisition section 1314, a provision content description information acquisition section 1315, and a recommendation method selection database 1317 are included.

  The switching KW content description information acquisition unit 1311 acquires content description information from the content description information management apparatus, and sends the content description keyword of the acquired content description information to the recommendation switching KW determination unit 1312a.

  The recommendation switching KW determination unit 1312a learns the user's interest and calculates a plurality of recommendation methods for calculating the evaluation value of the recommended content, the user ID, the content explanation keyword obtained from the switching KW content explanation information acquisition unit 1311, The selection is made based on the recommendation method selection database 1317.

  As shown in FIG. 24, the recommendation method selection database 1317 switches the recommendation method when the user ID 1317a for identifying the user and the content explanation information keyword acquired by the switching KW content explanation information acquisition unit 1311 are included. Information such as a switching KW 1317c, a switching recommendation method 1317d used when the content description information keyword includes the switching KW 1317c, and a normal recommendation method 1317b used when the content description information keyword does not include the switching KW 1317c is stored. ing.

  The content information recommendation unit is determined to be one of the recommendation methods selected by the content information selection unit 131b, and the content evaluation value is calculated from the recommendation content explanation information selected by the selection database 1310 and the user interest information 121. To do.

  The content information learning unit is determined from one of the recommended methods selected by the content information selection unit 131b, the learning content explanation information selected by the selection database 1310, and the user history obtained by the user history information input unit 135 From this, the user's interest information is learned.

Next, the operation of the information providing system 3 configured as described above will be described.
FIG. 25 is a flowchart showing an operation procedure of the information providing system 3 until the recommended content information is presented to the user. The operation of the information providing system 3 is basically the same as the operation until the recommended content of the information providing system 1 in Embodiment 1 is provided to the user. However, the step of calculating the content evaluation value (S704) is different from the first embodiment. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  First, a user sends a terminal ID and a user ID from the terminal 14 to the content recommendation device 13b via the network, and the content recommendation device 13b receives the terminal ID and the user ID at the terminal and the user ID input unit (S701).

  Subsequently, the content information selection unit 131b determines recommended content description information from the selection database 1310 using the terminal ID as a search key (S702). Subsequently, the content information selection unit 131b determines providing content description information from the selection database using the terminal ID as a search key (S703).

  The content description information is acquired in the switching KW content description information acquisition unit 1311 of the content information selection unit 131b. The content description information at this time may be determined by the content description information provider or may be determined for each user. The content description keyword of the acquired content description information is sent to the recommendation switching KW determination unit 1312a. The recommendation switching KW determination unit 1312a determines whether the user ID and the content explanation keyword are registered in the recommendation method selection database 1317 using the search key (S704a). When the content explanation keyword includes the switching KW, the recommendation method registered in the recommendation method for switching in the recommendation method selection database 1317 is selected, and the user ID is determined from the recommendation content explanation information and the user's interest information. The user interest information 121 is read from the user interest information management device 12 using the search key as a search key. Subsequently, the evaluation value of each content is calculated by the recommendation method selected earlier (S704b), and the recommended content is determined from the evaluation value. For example, in FIG. 24, when the user ID is “UID523” and the content explanation keyword includes “Japanese movie”, “recommendation method b” is selected as the recommendation method. Similarly, when the switching KW is not included in the content description keyword, a recommendation method registered in the normal recommendation method in the recommendation method selection database 1317 is selected, and the recommendation content description information and the user's interest information are used. The user interest information 121 is read from the user interest information management device 12 using the user ID as a search key. Subsequently, the evaluation value of each content is calculated by the recommendation method selected earlier (S704c), and the recommended content is determined from the evaluation value. For example, in FIG. 24, when the user ID is “UID 523” and the content explanation keyword does not include the switching KW, “recommended method a” is selected as the recommended method.

  Then, the recommended content providing unit 133 provides the content explanation information for provision corresponding to the recommended content determined by the content information recommendation unit to the user's terminal (S705).

  FIG. 26 is a flowchart showing an operation procedure of the information providing system 3 until learning of user interest information from the content explanation information and the user history. The operation of the information providing system 3 is basically the same as the operation of learning user interest information from the content explanation information and the user history of the information providing system 1 in the first embodiment. However, the step of calculating the evaluation value of the keyword from the user history and the content explanation information for learning is different from the first embodiment. Hereinafter, the same components as those in the first embodiment are denoted by the same reference numerals, description thereof is omitted, and different points will be mainly described.

  First, a user sends a terminal ID, a user ID, and a user history from the terminal 14 to the content recommendation device 13b via the network. The content recommendation device 13b is connected to the terminal and a user ID input unit (not shown). The terminal ID and the user ID are received, and the user history is received by the user history information input unit 135 (S901).

  Subsequently, the content information selection unit 131b determines the content explanation information for learning from the selection database 1310 using the terminal ID as a search key (S902).

  Next, the content description information is acquired by the switching KW content description information acquisition unit 1311 of the content information selection unit 131b. The content description information at this time may be determined by the content description information provider or may be determined for each user. The content description keyword of the acquired content description information is sent to the recommendation switching KW determination unit 1312a. The recommendation switching KW determination unit 1312a determines whether the user ID and the content explanation keyword are registered in the recommendation method selection database 1317 using the search key (S903a). The recommended method registered in the switching recommendation method of the recommendation method selection database 1317 is selected, and the selected content information learning unit calculates the evaluation value of the keyword from the learning content description information and the user history information ( S903b). Similarly, when the switching KW is not included in the content explanation keyword, the recommended method registered in the normal recommendation method of the recommendation method selection database 1317 is selected, and the selected content information learning unit A keyword evaluation value is calculated from the content description information and the user history information (S903c).

  Then, the content information learning unit 134 requests the user interest information management apparatus 12 to update the user interest information obtained from the user history input (S904).

  Thus, according to the information providing system 3 in the present embodiment, the user can obtain the same recommended content information regardless of the operated terminal and the browsed terminal, and depending on the content of the content explanation information The recommendation method can be changed. In other words, regardless of the content description information, the same content is recommended on any terminal, and all keywords are evaluated with the same weight, usually regardless of items such as program name and performer. The recommendation method is used, but when the keyword “movie” is included in the content description information, the recommendation method for calculating the evaluation value is changed by increasing the weight of the keyword indicating the performer over the other keywords. By doing so, it is possible to provide recommended content according to the user's preference.

  In this embodiment, a recommendation method for evaluating all keywords with the same weight and a recommendation method for evaluating with an increased performer weight are given as examples. However, the recommendation method is not limited to the above method. For example, a recommendation method for expanding a keyword input by a user in advance using a synonym dictionary, a recommendation method using only the genre information of a program provided by a content explanation information provider, and using a synonym dictionary using a synonym dictionary In addition, a recommendation method for creating new genre information and learning user interest in genre information, a recommendation method that introduces forgetting to gradually reduce the influence of past user history, and content description information are also taken into account Recommending method using collaborative filtering, recommending method to change user interest information depending on who sees, and recommending content by merging user interest information when receiving content recommendation by multiple users The recommendation method etc. which perform may be sufficient.

  In addition, in learning of user interest, there is a recommendation method for recommending a program using only the program selection history, a recommendation method for recommending a program using even the recording history of the program, not only interested, but also not interested. A program that recommends all user operation history as a user's history, such as a recommendation method for recommending programs using evaluations of user programs, such as certain or slightly uninterested, increasing the volume during content viewing, pausing, and resolution of viewing terminals A recommendation method, a recommendation method that uses the history of when and where the user watched the program, and a mobile phone that uses only the operation history obtained from the home TV by limiting the terminal that extracts the user's operation history. This is a recommendation method that uses only the operation history of a mobile phone that has personality. It may be.

  Note that the recommendation method selection database 1317 may be set by the user himself / herself or may be set by the content explanation information provider.

  The switching KW 1317c of the recommendation method selection database 1317 is a keyword, but may be a keyword logical expression expressed by logical sum or logical product, and is not limited to a keyword such as genre information of content.

  In this embodiment, the content explanation keyword is used as a switching criterion. However, the recommendation method may be switched depending on whether or not there is a certain amount of user history. In this case, the recommendation method is switched depending on whether there is a certain amount, but the recommendation method may be switched step by step. Further, the recommendation method may be switched depending on the frequency of access by a user such as a user who accesses frequently or a user who accesses only occasionally.

  In addition, the recommendation switching KW determination unit 1312a determines whether or not there is a keyword, but the content recommendation value is used in the normal recommendation method, and the switching recommendation method is used when the value is equal to or greater than a certain value. You may do it.

  In this embodiment, one recommendation method is used according to the switching keyword. However, a plurality of recommendation methods are used, and the ratio of programs recommended from each recommendation method is changed according to the user's history. May be.

  In addition, depending on the recommended method, the recommended content is displayed by distinguishing the recommended content, such as changing the color of the content description information to be provided or changing the font size. It may be possible to confirm whether or not.

  In addition, although the user interest information is one, the user interest information may be accumulated for each selected recommendation method.

  Also, by combining the present embodiment and the second embodiment, the content description information used for recommendation, learning and provision is changed or the recommendation method is changed according to the content description information. Is possible. The content description information used for recommendation and learning can be changed by the recommendation method.

(Embodiment 4)
In the first embodiment, the explanation information used for learning is predetermined according to the terminal as shown in FIG. However, with respect to highly public content such as program information, various information distributors provide content for explanation. On the other hand, in content recommendation based on a user's content processing history, recommendation accuracy may vary greatly depending on content description information.

  Therefore, in the present embodiment, when a plurality of pieces of program information exist, each program information is used to learn the user's preference, and which program information is obtained from the user's operation history with respect to the recommended content as a learning result. Is automatically determined to be suitable for the user, and program information used for program recommendation is selected. That is, a user's operations such as viewing and recording for each program are accumulated as user program operation information, and a plurality of recommended programs calculated using each program information are used to determine whether or not the user has viewed or recorded the recommended program. An apparatus for determining which program information is to be selected by the user and selecting a program to be browsed or recorded and determining a recommended program thereafter using the program information will be described. In the first embodiment, as shown in FIG. 5, program information for learning, program information for recommendation, and program information for provision are stored in advance according to the terminal on which the user views the program in advance. In the present embodiment, the table of the program information is automatically determined using the browsing history of the user's program.

  Note that the content described in the first to third embodiments is described as program content in the present embodiment as a specific example. Further, the content description information will be described as program information distributed in an electronic program or the like.

  An example of the system configuration of the information providing apparatus of this embodiment is shown in FIG. 27 includes a first program information storage unit 3101, a second program information storage unit 3102, a program information selection unit 3103, a user operation history information input unit 3104, a content information learning unit 3105, and user interest information storage. It comprises a section 3106, a program content recommendation section 3107, a recommended program content provision section 3108, a program content display section 3109, a program content display control section 3110, and a recommended program operation history storage section 3111. The operation of each module will be described.

  The first program information storage unit 3101 and the second program information storage unit 3102 store information related to each program distributed by broadcasting or the like. However, as shown in FIG. 28, the program information includes a plurality of explanation information for the same program, and is stored in the program information stored in the first program information storage unit and in the second program information storage unit. The program information being made is different. The example of FIG. 28 shows an example of simple program information (first program information) for mobile phones and standard program information (second program information) for televisions and PCs.

  The program information selection unit 3103 selects either the first program information stored in the first program information storage unit 3101 or the second program information stored in the second program information storage unit 3102. Use it to select whether to extract user preferences.

  The user operation history information input unit 3104 detects a user operation on the program content, and inputs information indicating the detected operation to the content information learning unit 3105 as an operation history for extracting user preferences. .

  The content information learning unit 3105 uses the operation history for each program input from the user operation history information input unit 3104 and the program information selected by the program information selection unit 3103 to determine the keyword of the program that matches the user's preference and its importance. Calculate the degree. An example is shown in FIG. Since the keywords that the user is interested in are extracted from the text information of the program information, when different program information is used, the keyword is naturally different from the keyword that is estimated to be of interest to the user. In particular, the types of keywords extracted using simple program information tend to decrease.

  The user interest information accumulating unit 3106 accumulates information related to keywords and their importance levels that match the user's preference calculated by the content information learning unit 3105. Information relating to the keywords and their weights shown in FIG. 29 is accumulated.

  The program content recommendation unit 3107 determines the program content to be recommended to the user by using the program information selected by the program information selection unit 3103 using the keyword stored in the user interest information storage unit 3106 and its importance. . FIG. 30 shows an example of a recommended program. As shown in FIG. 29, since the program information is different, the keywords estimated to be of interest to the user are different. Furthermore, because the keywords are different, the recommended programs are also different.

  The recommended program content providing unit 3108 provides the user with the program recommended by the program content recommendation unit 3107. FIG. 31 shows an example of providing recommended programs. In FIG. 31, a program recommended from the first program information is output first, and then a program recommended from the second program information is output. Either order may be sufficient as long as the program recommended from each program information is output.

  The program content display unit 3109 displays the program content provided by the recommended program content providing unit 3108 in accordance with a user request. For example, when the user checks the program information, a program output as shown in FIG. 31 as a result of inputting a recommended program with a remote controller or the like is output to a television screen or the like.

  The program content display control unit 3110 inputs a user's program content control instruction. For example, the user selects and browses a program, gives an instruction to record a program, and inputs an instruction to control the reproduction and recording of program content.

  In the program content display control unit 3110, the recommended program operation history storage unit 3111 stores operations such as recording and browsing related to the recommended program. Here, it is not the control related to the entire program content of the user, but the history of the user's input of operations such as browsing and recording regarding the program recommended by this system is accumulated. For example, as shown in FIG. 32A, the user has performed a browsing operation, recorded, or neither has been viewed or recorded for the recommended program (“×” in the “User Operation” column). ) Is accumulated.

  Further, in the program information selection unit 3103, a table (FIG. 32 (b)) relating to user operation weights for programs recommended in advance is accumulated, and an evaluation value can be input in accordance with user operations for the recommended programs. It is like that. Further, an evaluation value is input for each program in the table created as shown in FIG. 32A, and the sum of the evaluation values is calculated for each program information. Using the result (FIG. 32 (c)), program information suitable for determining a recommended program is determined. In the case of FIG. 32 (c), it is determined that the program recommended by using the second program information is often viewed or recorded by the user, so the second program information is used. As described above, the program information selection unit 3103 performs selection.

  The operation of the information providing apparatus configured as described above will be described using the flow of FIG. First, the program content recommendation unit 3107 determines whether or not a user has requested output of a recommended program (S3801). If there is a request from the user, the process proceeds to the next step. If there is no request, this step is repeated. When there is a program recommendation request from the user, first, using the first program information stored in the first program information storage unit 3101, the keyword stored in the user interest information storage unit 3106 Using the weight, the program content recommendation unit 3107 determines a recommended program (S3802). As for the recommended program determination method, when a keyword is extracted from the text of program information that the user has browsed or recorded in the past in the same manner as in the first embodiment, and a program including the keyword is broadcast in the future. Outputs the program as a recommended program. Similarly, using the second program information stored in the second program information storage unit 3102, the keyword stored in the user interest information storage unit 3106 and its weight are used to program content recommendation unit 3107. In step S3803, a recommended program is determined. As a result, recommended programs as shown in FIG. 30 are determined, respectively. As shown in FIG. 31, the recommended program content providing unit 3108 uses the program content mixed with each recommended program as a recommended program, and a program content display unit 3109. (S3804).

  Next, this step is repeated until all the program contents presented in S3804 are broadcast (S3805). When all the recommended programs are broadcast, the recommended program operation history accumulating unit 3111 stores what operations such as browsing and recording have been performed on each recommended program. The table shown in FIG. 32A is created using the operation history (S3806). For each program information used for recommendation, the sum of the evaluation values of the recommended programs is calculated by the program information selection unit 3103 as shown in FIG. 32C (S3807). The program information with the largest sum of results is used as program information for determining a future recommended program (S3808).

  As a result of the above operation, the recommended program was initially determined using both the first program information and the second program information, but thereafter, the recommended program is determined using the second program information. It will be. However, again, it may be suitable for the user's operation to determine the recommended program using both the first program information and the second program information. Therefore, although the recommendation result using the first program information is not presented to the user, the learning result in the content information learning unit 3105 is used as the first result in the user interest information storage unit 3106 using the first program information. It is necessary to accumulate as user interest information using the program information.

  As a result of the above operation, the recommended program was initially determined using both the first program information and the second program information, but thereafter, the recommended program is determined using the second program information. It will be. The recommended program is determined using the second program information, but this method is applied again when the user no longer performs operations such as recording and browsing for the program recommended by the second program information. In the flow of the present embodiment, in the situation where the recommended program is determined using both the first program information and the second program information, either the first program information or the second program information is used. In the situation where the recommended program is determined by the second program information, the first program information or both the first program information and the second program information is used. It may be determined whether to use.

  As a result of the above operation, when there is a plurality of program information, even if there is a difference in the programs recommended by the respective program information, the optimum program recommendation can be made by using the user operation history for the recommended programs. Program information that can be performed can be automatically selected.

  In the fourth embodiment, both the recommended program based on the first program information and the recommended program based on the second program information are presented and used for recommendation depending on which program is viewed and recorded by the user. The program information was selected. However, although the recommendation result is presented by the first program information, the user may record the program at his own will. Therefore, in the fourth embodiment, not only the operation history of the program recommended by the first program information but also the operation history of the program recorded or browsed at the user's will is used to select program information used for recommendation. A device according to a modification will be described.

  FIG. 34 shows the system configuration of this modification. 34 includes a first program information storage unit 3901, a second program information storage unit 3902, a program information selection unit 3903, a user operation history information input unit 3904, a content information learning unit 3905, and a user interest information storage. A section 3906, a program content recommendation section 3907, a recommended program content provision section 3908, a program content display section 3909, a program content display control section 3910, and a program recording operation history storage section 3911. The operation of each module will be described. However, the modules other than the program recording operation history accumulation unit 3911 perform the same processing as in the fourth embodiment, and thus description thereof is omitted.

  FIG. 35 shows a program recommended by the program content recommendation unit 3907 using the program information stored in the first program information storage unit 3901. When there is a request for recommended program information from the user, the program shown in FIG. 35 is output and provided.

  The program recording operation history storage unit 3911 stores program information recorded by the user. Unlike the fourth embodiment, this recording operation history stores not only the recording operation history of recommended programs but also the history of program recordings entered by the user at his / her own will. For example, as shown in FIG. 36A, eight programs from P1002 to P1108 are recorded by the user. Among them, the three programs P1002, P1005, and P1312 are recommended programs using the program information 1. However, when the recommended program is determined using the program information 2, it is possible to recommend six programs P1002, P1005, P1102, P1302, P1312, and P1108.

The program information selection unit 3903 uses the table of FIG. 36A to calculate, using statistics, whether the program information 1 or the program information 2 is more appropriate for the user. To do. For example, as shown in FIG. 36 (b), the recommended program recall rate when each piece of program information is used is calculated. However, the recommended program recall is
Recommended program recall =
(Number of recorded programs among programs recommended using predetermined program information) / (Number of programs recommended using predetermined program information)
Calculated by In the case of FIG. 36A, it is determined that the program information 2 is used to reproduce the program recorded by the user, and program recommendation is performed using the program information 2 in the future.

  The above operation will be described with reference to the flowchart of FIG. First, the program content recommendation unit 3907 determines whether or not there is a request for outputting a recommended program from the user (S4201). If there is a request from the user, the process proceeds to the next step. If there is no request, this step is repeated. When there is a program recommendation request from the user, first, using the first program information stored in the first program information storage unit 3901, the keyword stored in the user interest information storage unit 3906 Using the weight, a recommended program is determined and presented by the program content recommendation unit 3907 (S4202). Similarly, using the second program information stored in the second program information storage unit 3902 and using the keyword stored in the user interest information storage unit 3906 and its weight, the program content recommendation unit 3907 is used. In step S4203, a recommended program is determined.

  Next, this step is repeated until all the program contents presented in S4202 are broadcast (S4204). When all recommended programs are aired, since the program recording operation history storage unit 3911 stores what operations such as browsing and recording have been performed on each recommended program, the program information selection unit 3903 The table shown in FIG. 36A is created using the operation history (S4205). For each program information used for recommendation, the recorded program reproduction rate is calculated as shown in FIG. 36B (S4206). As a result, program information having the largest recorded program reproduction rate is selected as program recommendation program information. In this example, the program information 1 is changed to the program information 2 (S4207).

  As a result of the above operation, the recommended program is determined based on the program information 1, but the recommended program is determined based on the program information 2 thereafter. However, although it changed to the program information 2 by S4207, the program recommended continuously with the program information 1 may be determined depending on the value of the program reproduction rate. Therefore, it may be automatically determined whether the program information 1 or the program information 2 should be used by periodically executing this step. Thereafter, when there is a possibility that program information may be changed, the content information learning unit always learns the user's interest information in the user interest information storage unit without presenting the recommended program to the user. Need to be updated.

  As a result of the above operation, if there is a plurality of program information, even if there is a difference in the programs recommended by each program information, the program information that can reproduce the recording history most using the user's recording history By selecting, it is possible to determine program information for which an optimal program recommendation can be made.

  In the fourth embodiment, one content information learning unit is used to make use of the difference between the recommended program using the first program information and the recommended program using the second program information. A system for automatically determining whether the program information is suitable for the user has been described. However, even when there are a plurality of content information learning units using the same program information, a plurality of types of recommended programs are output. This system configuration is shown in FIG. In FIG. 38, there are a first program recommendation service 4303 and a second program recommendation service 4304, and even if the same user operation history is used, different programs are recommended for each program recommendation. Therefore, by using the recording history of the user's program, it is possible to automatically determine which program recommendation output is suitable.

  In the system configuration of FIG. 38, the recommended program content control unit 4306 appropriately selects the recommended program output from each program content information recommendation unit 4305. As the selection method, the method described in the fourth embodiment or the modification of the fourth embodiment is used. For example, as shown in FIG. 39, a recommended program from each program content information recommendation unit 4305 is received. However, as for the recommended program output from each program content information recommendation unit 4305, program information on a plurality of channels may be determined as a recommended program in the same time zone. When only one program can be received at the same time in an apparatus for recording a program, a plurality of programs cannot be recorded simultaneously. Therefore, regarding these programs, it is determined whether the user's preference can be expressed by determining which program information of the first and second program information is used to determine the recommended program. The program to be preferentially presented is determined using.

  As a result of the above operation, when there are a plurality of program recommendation services, even if there is a difference in the programs recommended by the respective program information, the recommended operation is synthesized by using the user operation history and an appropriate The recommended program can be presented to the user.

  In the present embodiment, as shown in FIG. 27, the information providing server performs program information selection processing. However, only the first program information and the second program information are accumulated in the server, The process may be executed at the terminal. Furthermore, a plurality of program information may be received by broadcasting or the like, and the program selection process may be performed only by the terminal.

(Embodiment 5)
In the fourth embodiment, the method of selecting program information suitable for program recommendation using the operation history of the user's program has been described. As a result, program information suitable for user program recommendation can be selected. However, in each program information, one program information may be more suitable for a specific genre, or one program information may be more suitable for a specific time zone.

  Therefore, in the present embodiment, not only program information is selected, but also recommended conditions such as a specific genre, time zone, and channel are extracted, and the recommendation result from predetermined program information is given priority in the recommended conditions. An information providing apparatus that outputs automatically will be described.

  FIG. 40 shows a system configuration of the information providing apparatus according to the present embodiment. 40 includes a first program information storage unit 4501, a second program information storage unit 4502, a user history information input unit 4503, a content information learning unit 4504, a user interest information storage unit 4505, and a program content recommendation unit. 4506, a recommended program content control providing unit 4507, a program content display unit 4508, a program content display control unit 4509, a recommended program operation history storage unit 4510, and a recommended condition determination unit 4511. Next, the operation of each module will be described. However, the operations of the constituent elements 4501 to 4506, 4508, and 4509 are the same as those of the constituent elements having the same names in the fourth embodiment, and thus description thereof is omitted.

  The recommended program content control providing unit 4507 is stored in the first program information storage unit 4501 by using the keyword expressing the user's interest (preference) stored in the user interest information storage unit 4505 and information on its weight. The program content to be provided to the user is determined from the recommended program based on the current program information and the recommended program based on the program information stored in the second program information storage unit 4502. For example, as shown in FIG. 41, a recommended program (FIG. 41 (a)) using the first program information and a recommended program (FIG. 41 (b)) using the second program information are input. Here, the commonly recommended programs P1002, P1005, P1312, and P1210 are first set as recommended programs. Next, the upper recommended program output from each program information is set as the final recommended program. As a result, in the example of FIG. 41, 10 programs shown in FIG. 41C are determined as recommended programs. In both program information 1 and program information 2, seven programs out of the top nine recommended programs are selected.

  The recommended program operation history accumulation unit 4510 accumulates the user operation history of the program provided to the user as a recommended program by the recommended program content control providing unit 4507. Specifically, the presence / absence of user operation such as whether the recommended program has been viewed or recorded by the user is stored. For example, as shown in FIG. 42, information such as the broadcast time zone, channel, genre, day of the week, etc. of each program and program information used to recommend the program are stored. Whether or not the user has browsed or recorded (“◯” in the “user operation” column) or not (“×”) is stored.

  The recommended condition determination unit 4511 determines conditions to be preferentially presented to the user as a result of recommendation using each program information from the history accumulated in the recommended program operation history accumulation unit 4510. Specifically, for a recommended program, a condition common to the program operated by the user is extracted, and a program that satisfies the condition is surely output as a recommended program from the next time. In this embodiment, program information for recommendation is selected in the fourth embodiment, but in this system, a recommendation condition is extracted for each program information, and a program satisfying the recommendation condition is recommended. In this case, the recommended program is preferentially presented to the user.

  In general, the degree of detail of program information varies depending on the provider, genre, time zone, day of the week, and channel. If it is described in detail, the possibility of being output as a recommended program is increased depending on the amount of text information. On the other hand, when there are few descriptions, it is difficult to output as a recommended program. Therefore, by extracting conditions such as a genre and a channel preferentially recommended for each program information, it becomes possible to perform highly accurate recommendation suitable for the user. FIG. 43 shows the presence / absence of a user operation for each piece of program information.

Further, with respect to the program information 1 in FIG. 43 (a), as shown in FIG. 44, the number of programs is calculated for each attribute and its attribute value, and each attribute value was browsed or recorded. The number of programs and the number of programs that have not been viewed or recorded are calculated. Next, for each attribute value, the user browsing recording rate User browsing recording rate =
(Number of programs recorded or viewed in recommended programs with predetermined attributes) / (Number of recommended programs with predetermined attributes)
Calculate with The fact that this value is close to 1.0 indicates that there is a high possibility that a program recommended under that condition is viewed or recorded. Therefore, a condition in which the user browsing recording rate is larger than a predetermined value (for example, 0.95) is determined as a recommendation condition for preferentially recommending a program. In the example of FIG. 44, when the recommended program is determined using the program information 1, all the programs with “time zone = midnight” are browsed or recorded by the user, so the recommended program with “time zone = midnight”. Is preferentially output as a recommended program.

  Based on the recommendation condition determined by the recommendation condition determination unit 4511, the recommended program content control providing unit 4507 preferentially presents a program satisfying the recommendation condition to the user as a recommended program. For example, as shown in FIG. 45, when there is a table of recommended programs using program information 1 (FIG. 45 (a)) and a table of recommended programs using program information 2 (FIG. 45 (b)), In each piece of program information, a program that satisfies the recommended condition determined by the recommended condition determining unit 4511 is set as a recommended program that is preferentially presented to the user. In the recommended program list using the program information 1, since P1005 and P1201 match the recommended conditions, they are recommended programs. In particular, regarding P1201, the recommendation ranking is low, but it is a recommended program. In the recommended program list using the program information 2, since P1120 matches the recommended condition, it becomes a recommended program. Finally, a table of recommended programs as shown in (c) is generated.

  By performing the operations described above, the user's operation history information for the recommended program is used to automatically extract preferentially recommended conditions (genre, time zone, channel, day of the week, etc.) It is possible to output recommended programs with high accuracy.

  The above operation will be described centering on the operation of the recommendation condition determination unit 4511 with reference to the flowchart of FIG. The user operation history for the recommended program stored in the recommended program operation history storage unit 4510 is extracted as shown in FIG. 42 (S5101). Next, the recommendation condition determination unit 4511 creates a table shown in FIG. 43 for each piece of program information used for recommendation (S5102). First, program information 1 is selected (S5103). As shown in FIG. 44, the number of programs is calculated for each attribute of the program and its attribute value (S5104). The user browsing recording rate is calculated for each attribute value (S5105). It is determined whether there is an attribute having a user browsing recording rate of 1.0 (S5106). In the example of FIG. 44, since the viewing recording rate is 1.0 in the case of the program “time zone = midnight”, the attribute value is set as the recommendation condition (S5107). If there is no attribute value with a viewing recording rate of 1.0, another program information (program information 2) is used to calculate again (S5108, S5109).

  This step is repeated until the recommended program is output from the program content recommendation unit 4506 to the recommended program content control providing unit 4507 (S5110). If it is output, it is determined whether there is a program having a recommendation condition for each program information (S5111). In this example, since programs P1005 and P1201 satisfying the recommendation conditions exist in FIG. 45, the programs are set as recommended programs (S5112). A recommended program is determined by adding a program that is recommended in a higher rank among other programs (S5113). As a result, the recommended program shown in FIG. 45C is determined. If there is no recommendation condition, a recommended program is determined from the programs recommended in the higher rank in each program information (S5114).

  As a result of the above operation, when there is a plurality of program information, even if there is a difference in the programs recommended for each program information, the user's recording history is used to preferentially recommend each program information. It is possible to automatically determine the conditions for the program and realize highly accurate program recommendation.

(Embodiment 6)
In the fourth embodiment and the fifth embodiment, the method of realizing highly accurate program recommendation by selectively using program information when performing program recommendation has been described. In particular, in the fifth embodiment, with respect to predetermined program information, program recommendation is realized preferentially under specific conditions (genre etc.) with respect to the result of program recommendation.

  However, when the user's operation history is not sufficiently accumulated, there are cases where sufficient preferences cannot be extracted even using program information distributed from a specific trader. In particular, unless the program information is described in detailed text, keywords that are of interest to the user cannot often be extracted sufficiently. Therefore, in the present embodiment, a program recommendation system is provided in which a plurality of pieces of program information are combined to extract a keyword expressing the user's preference even when there are few programs selected by the user.

  FIG. 47 shows the system configuration of the information providing apparatus of this embodiment. 47 includes a first program information storage unit 5201, a second program information storage unit 5202, a program information synthesis unit 5203, a program information selection unit 5204, a user operation history information input unit 5205, and a content information learning unit. 5206, a user interest information storage unit 5207, a program content recommendation unit 5208, a recommended program content provision unit 5209, a program content display unit 5210, a program content display control unit 5211, and a recommended program operation history storage unit 5212. The operation of each module will be described. Note that operations from 5201, 5202, 5204 to 5211 are the same as those in the fourth embodiment, and thus the description thereof is omitted. In this embodiment, only program information is acquired through a network or broadcast wave, and the user's interest information or program recommendation operation module is operated on a home recording device or receiver.

  The program information combining unit 5203 combines the program information stored in the first program information storage unit 5201 and the program information stored in the second program information storage unit 5202 to generate new program information. . For example, as shown in FIG. 48, new program information is generated by synthesizing the contents of program information determined to be the same program based on the broadcast time and channel.

  In the program content recommendation unit 5208, the recommended programs are determined using the synthesized program information, so that eight recommended programs from program A to program H can be determined as shown in FIG.

  The recommended program operation history storage unit 5212 stores a user operation history for the program recommended by the recommended program content providing unit 5209. As shown in FIG. 49, it is accumulated that the programs viewed or recorded by the user for the recommended programs A to H are the four programs B, D, E, and F.

  The program information selection unit 5204 is used for program recommendation from the program information of the first program information storage unit 5201, the program information of the second program information storage unit 5202, or the program information synthesized by the program information synthesis unit 5203. Program information is selected using the operation history for the recommended program stored in the recommended program operation history storage unit 5212. As shown in FIG. 49, when the program recommendation is performed using the synthesized program information, the programs A to H are recommended, but there are four programs viewed or recorded by the user. On the other hand, when program recommendation is performed using program information 1, five programs A, B, C, E, and G are recommended. When the program information 2 is used, six programs B, C, D, E, F, and H are recommended.

Therefore, the program information selection unit 5204 reproduces the program operated by the user.
(Number of recommended programs that have been operated by the user in programs recommended by the predetermined program information) / (Number of recommended programs that have been operated by the user)
, And recommend from the next time using program information higher than a predetermined value.

  In the example of FIG. 49, when program recommendation is performed using the program information 2, it can be seen that all the programs viewed or recorded by the user are among the recommended programs. Therefore, from the next time, program recommendation is performed using the program information 2.

  The above operation will be described with reference to the flowchart of FIG. First, the program information synthesis unit 5203 synthesizes the program information stored in the first program information storage unit 5201 and the program information stored in the second program information storage unit 5202 to obtain new program information. Generate (S5501). This step is repeated until the user requests program recommendation (S5502). When the program request is requested by the user, the program content synthesizing unit 5203 uses the keyword estimated by the program content recommendation unit 5208 that is accumulated in the user interest information accumulating unit 5207 to be synthesized by the program information synthesizing unit 5203. The recommended program is selected using the program information thus obtained (S5503). The recommended program content providing unit 5209 presents these program contents (program A to program H in FIG. 49) to the user via the program content display unit 5210, and this step is performed until all the recommended programs are broadcast. Is repeated (S5504).

  When all the programs recommended by the program information synthesized by the program information synthesizing unit 5203 are broadcast, the program content recommendation unit 5208 reads the recommended program from the operation history accumulated in the recommended program operation history accumulation unit 5212. The operation history of the user is extracted (S5504). For each recommended program, in this example, programs B, D, and H were recorded or viewed as shown in FIG. Next, the program content recommendation unit 5208 is stored not in the program information combined by the program information combining unit 5203 but in each storage unit of the first program information storage unit 5201 and the second program information storage unit 5202. A recommended program is determined using only the program information (S5506, S5507). When program information 1 is used, programs A, B, C, E, G, and program information 2 are used, and programs B, C, D, E, F, and H are calculated as recommended programs. . The program information selection unit 5204 calculates a program reproduction rate when the program is recommended using each program information (S5508). As a result, in the case of this example, as shown in FIG. 49, the recall rate when each program information is used is calculated, and the recommendation result based on the program information 2 is the recall rate 1.0. Go ahead. Since the reproduction rate based on the program information 2 is 1.0, in the next program recommendation, program recommendation using the program information 2 is performed instead of synthesizing the program information (S5509). If there is no program information with a recall ratio of 1.0, the synthesized program information is used after the next time (S5510).

  As a result of the above operation, when there are a plurality of pieces of program information, it is possible to generate new program information by synthesizing those pieces of program information and to recommend as much program information as possible to the user. Furthermore, by using the history that the user has viewed and recorded from the recommended programs, it becomes possible to automatically select program information when providing recommended programs to the user.

  In this embodiment, the reproduction rate of the recommended program is used for selecting the program information by the program information selection unit 5204. However, the relevance rate of the recommended program may be used. In the present embodiment, program information having a recall ratio of 1.0 is selected, but selection of program information having the highest recall ratio or selection based on a predetermined threshold may be used.

  In the present embodiment, the program information is generated by synthesizing two programs of program information 1 and program information 2. However, when there are three or more program information, a combination of the program information is selected. To select program information.

  In the present embodiment, new program information is generated by combining program information, and a recommended program is determined. However, it is also possible to determine each recommended program using each program information, and to determine a wide range of recommended programs by combining the recommended programs. However, it is possible to determine a wider range of programs by determining recommended programs using new program information obtained by combining program information. This is because keywords are extracted in various ways when extracting personal interest information, and a lot of keywords are included in determining a recommended program.

(Embodiment 7)
In the sixth embodiment, the recommendation results of a wide range of programs can be obtained by synthesizing program information. However, when program information obtained by synthesizing a plurality of program information is always used, a program that does not match the user's preference is also recommended. Therefore, program information can be recommended so that only program information that meets the user's preference can be recommended. He also described how to choose.

  In particular, synthesizing program information is effective when the number of operation histories from the user is not sufficient and a wide range of program recommendations are required. On the other hand, when the number of user operation histories is not sufficient, appropriate recommendation cannot be performed. For example, the user's preference for the program is most expressed in the recording operation, and in the browsing operation, it is not always possible to accurately detect whether the user is browsing the program. Furthermore, even in the recording operation, in the case of a program that has only been recorded but is not being browsed, the user's preference may not be expressed sufficiently by the recording operation alone. On the other hand, if a program is recommended using only the history of viewing a recorded program, an accurate program recommendation may not be performed at the initial stage of learning because a sufficient operation history cannot be obtained.

  Therefore, in the present embodiment, an accurate program recommendation is realized by selectively using information of a user's operation history used for program recommendation. FIG. 51 shows the system configuration of the information providing apparatus of this embodiment. 51 includes a program information storage unit 5601, a user operation history information input unit 5602, a content information learning unit 5603, a user interest information storage unit 5604, a program content information recommendation unit 5605, a recommended program content provision unit 5606, and a program. A content display unit 5607, a program content display control unit 5608, a user operation history storage unit 5609, a recommended program operation history storage unit 5610, and a learning operation history selection unit 5611 are included. Next, the operation of each module will be described. Note that the components 5601 to 5608 perform the same processing as the components having the same names in the fourth embodiment, and thus the description thereof is omitted.

  The user operation history accumulation unit 5609 accumulates operation information for the user program content. Note that the user's operation on the program content is generally operated from a plurality of devices. For example, the program content may be browsed on a television, or the program may be recorded on a recording device, and the program content may be played back on the device. Furthermore, recording reservations may be made from outside using a mobile phone or the like, and program content recorded on the mobile phone may be played. Therefore, FIG. 52 shows a system configuration in which a user's operations on program content are accumulated as a history in devices connected via a network. FIG. 52 includes a user operation history storage unit 5701, user operation input units 5702, 5705, 5708, and 5711, device control units 5703, 5707, 5710, and 5712, a content display unit 5704, and content storage units 5706 and 5709. .

  The user operation history storage unit 5701 stores control contents of program content in each content control device (digital television, DVD recorder, hard disk recorder, mobile phone, etc.). In each device, the user inputs control of the program content via the user operation device input unit, and each device control unit operates according to the input. Device operation information is stored in the user operation history storage unit via the network.

  In the user operation history storage unit 5609 (5701), as shown in FIG. 53, the operation history related to the program content in each device is stored as the operation date and time, the operation details, and the program content information as the operation target. In addition, among the operations of the program content, operations that reflect the user's preference include program browsing, recording, and playback of recorded programs. Therefore, information related to these operations and program content that is the target of the operations is extracted and used as an operation history for learning user preferences.

  The definition of the operation history for learning and the number of program information related to the operation history are shown in FIG. As shown in FIG. 54, the operation of browsing a program means that the program content is browsed when there is a history of the program being displayed on the television screen for 80% or more of the broadcast time of the program. become. The operation of recording a program means that the program content is recorded when there is a history of performing the recording operation in advance or during the broadcast of the program. Furthermore, (viewing history) ∪ (recording history) means an operation of a program that has been viewed or recorded, and the program content that is the target is program content that is used for learning. As a result, as shown in FIG. 54, the number of programs targeted for each operation is calculated, and the selected programs differ depending on the type of operation history. As programs used for learning differ, program recommendation results also differ.

  The recommended program operation history storage unit 5610 stores a history of whether or not the recommended program has been recorded or viewed by the user. The quality of the recommended program is judged according to the contents of the operation.

  The learning operation history selection unit 5611 selects a user operation history used for extracting keywords and weights of programs that the user likes. For example, when the user's preference is learned using a program from which the user's browsing or recording history is obtained, and when the user's preference is learned using a program from which the user's browsing history is obtained, Keywords that express preferences are different. In particular, in the initial stage of learning user preferences (when the number of programs to learn is small), the results differ greatly depending on which program content is used to extract user preferences. Therefore, in the present embodiment, at the initial stage of learning, in order to perform learning using as many program contents as possible, it is determined that the program content is a program content that the user prefers to view or record. Perform preference learning. For example, when learning is performed using program content (60 programs) with a user's browsing or recording history, it is assumed that eight programs from program A to program H are recommended as shown in FIG. However, it is assumed that there are five programs B, C, D, E, and F that are viewed or recorded by the user with respect to the recommended programs. At this time, when the program is recommended using only the program content (52 programs) having the browsing history without using the recording history as the user operation history, the programs B, C, D, E, F, 5 programs of G are recommended. Similarly, when program recommendation is performed using only program content with a recording history (21 programs), there are four programs B, D, F, and G. Similarly, when program recommendation is performed using only program content (12 programs) having a history of (recording + playback), there are two programs B and D.

Therefore, the effectiveness of the user operation history used for the program recommendation is evaluated. For example, as used in the fifth embodiment, the evaluation is performed using the reproduction rate of the program. However, recall is
Reproducibility =
(Number of recommended programs that can be recommended using a predetermined history among programs recorded or browsed by the user) / (Number of recommended programs that the user has recorded or browsed among programs recommended by the system)
It is. This value is calculated in each operation history of the user, and if there is a history with a recall ratio of 1.0, a program is recommended using the history in the future. In the example of FIG. 55, when the program recommendation is performed using the browsing history, the recall rate is 1.0. Therefore, it is assumed that program recommendation will be performed using this history in the future.

  The above operation will be described with reference to the flowchart of FIG. This step is repeated until a program recommendation request from the user is accepted (S6101). When the program recommendation request is received, the program content information recommendation unit 5605 first determines a recommended program using the program content having the (viewing + recording) operation history (S6102). As a result, when 60 programs are used as shown in FIG. 55, programs A to H are determined as recommended programs. A program recommendation result is calculated using the operation history of the browsing-only history, the recording-only history, and the (recording + playback) history (S6103). As a result, it can be seen that a part of programs A to H is recommended as shown in FIG.

  This step is repeated until all recommended programs are broadcast using the (browsing + recording) history (S6104). When everything is broadcast, the user's operation history for the recommended program is extracted from the recommended program operation history storage unit 5610 via the learning operation history selection unit 5611 (S6105). As a result, as shown in FIG. 55, the five programs from program B to program F are viewed or recorded by the user. Using the recommended program using each history, the program content information recommendation unit 5605 calculates the reproducibility of program recommendation for each history (S6106). As a result, the recall for each history is calculated as shown in FIG. It is determined whether there is a recommendation result having a recall ratio of 1.0 using each history (S6107). In the case of this example, the result of program recommendation using the browsing history has a recall rate of 1.0. The next program recommendation is a program recommendation using only a browsing history with a recall rate of 1.0 (S6108). If there is no history with a recall ratio of 1.0, program recommendation is performed again using the (viewing + recording) history again.

  As a result of the above operation, at the initial stage of learning, it is possible to recommend programs using program information that has been operated by the user, and to recommend as much program information as possible. Furthermore, it is possible to automatically select an operation history to be used for recommending a program by using a user's browsing / operation history of the recommended program.

  As mentioned above, although the information provision apparatus which concerns on this invention was demonstrated based on Embodiment 1-7 and its modification, this invention is not limited to these Embodiment and modification. For example, embodiments that are realized by arbitrarily combining the constituent elements in each embodiment within a consistent range, and those in which those skilled in the art have conceived modifications are also included in the present invention.

  Further, as a modification of the first embodiment, the following form may be adopted. That is, in the first embodiment, as shown in FIG. 5, the program guide to be displayed is determined in advance according to the terminal, and the program guide corresponding to the terminal to be displayed is displayed. However, in a terminal that displays a program guide, functions used in each terminal may be determined. In particular, if the user's intention to display the program guide is clear, the program guide should be displayed accordingly. For example, when a program guide is displayed to check whether there is a recommended program among currently broadcast programs on a television terminal in the home, the program guide for one week ahead is displayed. There is no need to display. On the other hand, if the mobile phone is not equipped with a tuner for receiving broadcasts, it may be meaningless to display the currently broadcast program guide. In the case of a mobile phone, when the user mainly performs remote reserved recording, the user may only need to display information of a recommended program to be reserved for recording, instead of the currently broadcasted program.

  Therefore, in the present embodiment, when a terminal that displays a program guide is determined, a history of what operations the user has performed in the terminal that displays the program guide in the past is accumulated. Is selected and displayed. FIG. 57 shows the system configuration.

  In FIG. 57, the same numbers are assigned to portions that perform the same processing as in the first embodiment. As modules different from the first embodiment, there are a device specification information storage unit 6201, a program guide display specification storage unit 6202, and a program guide browsing intention estimation unit 6203. The operation of the newly added part will be described next.

  The device specification information storage unit 6201 stores functions that can be used by the user in the terminal that displays the program guide. In recent years, mobile phones have become highly functional, and it has become possible to browse broadcast contents. Also, with regard to PCs, programs can be viewed and recorded. However, not all models have the same functions. Thus, for example, as shown in FIG. 58, what functions are provided for program information in a terminal capable of displaying a program table owned by the user is stored in advance. The table in FIG. 58 may be automatically registered by each device via a network. Also, only the functions used by the user may be registered manually.

  The program guide display specification storage unit 6202 stores specifications for displaying a program guide when the user's purpose of displaying the program guide can be estimated in a terminal that displays the program guide. For example, as shown in FIG. 59, “browsing the program guide” is a more detailed information on each program than when viewing the program guide to confirm what kind of broadcast this week is available in general broadcasting. It is better to display the entire program table as much as possible and easily check whether there are programs that are being browsed regularly this week. Also, when displaying program information for viewing program content, it is desirable to determine whether there is a recommended program in the currently broadcast program rather than displaying a future program. Therefore, detailed information on programs currently being broadcast should be displayed as a program guide. In addition, if only scheduled recording is performed by looking at the program guide, whether or not to record the program by displaying a list of recommended programs and displaying the contents instead of displaying the entire program guide. It becomes possible for the user to determine whether or not. As described above, when the program guide is displayed, the optimum program guide display form differs depending on the application (confirmation of the program guide, browsing of program contents, reserved recording, etc.).

  The program guide browsing intention estimation unit 6203 estimates the purpose of displaying the program guide at each terminal from the user's operation history when the program guide is displayed. For example, as shown in FIG. 60 (a), after displaying the program guide, a history of what operation the user has performed on which terminal is accumulated in the user operation history information input unit, and the history information is used. Thus, the purpose for the user to display the program guide is estimated. From the history, for example, since the tuner is not installed from the function of FIG. 58 in the mobile phone, the program content currently being broadcast cannot be browsed. However, it is only possible to perform scheduled recording for a recording device in the home and browsing of a future program guide. In the example of FIG. 60A, it is assumed that, from the browsing operation history of the user's program guide, the program guide is not browsed and only the reserved recording of the program is performed. When the operation history for such a program guide has been accumulated, as shown in FIG. 60 (b), it is expected that when the user views the program guide with a mobile phone, the reserved recording will be performed. The Therefore, as shown in FIG. 59, when a program guide is displayed for the purpose of reserved program recording, it is determined that only information related to a recommended program should be displayed using program explanation information for television.

  As a result, when the user makes a program recommendation request, a program table for a mobile phone is usually displayed as shown in FIG. 61 (a). In the case of the present invention, FIG. 61 (b) The recommended program and its detailed information are displayed as shown in FIG.

  As a result of the above operation, it is possible to display an appropriate program guide according to the purpose of use of the program guide at each terminal using the operation history of the program guide at each terminal.

  The above operation is summarized in the flow of FIG. In step S6701, this step is repeated until a program recommendation request is received, as in the fourth embodiment. If there is a program recommendation request, it is detected from which terminal a request to display the program guide and the recommended program is made (S6702). In the case of the present embodiment, it is assumed that there is a program table display request from a mobile phone. Next, in the program guide browsing intention estimation unit 6203, what operation was performed when the program guide was displayed in the past week is extracted from the user history information input unit (S6703). In the present embodiment, it is assumed that the history information shown in FIG. It is determined whether or not the intention of displaying the program guide can be specified from the operation history (S6704). In the present embodiment, as shown in FIG. 60 (b), in the history information, only a program recording reservation is made on the mobile phone. Therefore, when there is a program recommendation request from the mobile phone, it is presumed that the reserved recording of the program is the purpose again. Since the purpose has been estimated in this way, the process proceeds to S6705. The recommended content providing unit determines and outputs the recommended content provision specification based on the result estimated by the program guide browsing intention estimation unit 6203 and the program guide display specification stored in the program guide display specification storage unit 6202. (S6705). In this embodiment, since it is determined that the program table is displayed for the purpose of recording the program on the mobile phone, the recommended program as shown in FIG. 61 (b) according to the contents of FIG. 59 stored in the program guide display specification storage unit 6202. Is displayed. If the display intention of the program guide cannot be estimated in S6704, the program guide is displayed according to the preset table of FIG. 5 as in the first embodiment (S6706).

  As a result of the above operation, when viewing the program guide in the usual way of using the program guide of the user, the program guide can be displayed with suitable display specifications when the program guide is displayed.

  It is also possible to select a period of a recommended program candidate for each terminal that has requested display of a recommended program. For example, when the program guide browsing intention estimation unit 6203 can determine whether the user's intention to browse the program is recording or program viewing, and if the recording is intended, the program recommended from the program guide up to one week ahead is selected. When the selected program is to browse the program, it is possible to determine a recommended program from currently broadcast programs. That is, it becomes possible to select a recommended period of a program according to the intention of browsing the program guide.

  In the present embodiment, the display purpose of the program guide is estimated for each terminal, but the display form of the program guide may be changed for each time zone in which the terminal is used. For example, in the morning (before going to work), the user browses the program guide and performs scheduled recording in the morning (before work), but only browses the program guide after returning home. In that case, it can be realized by changing the display format of the morning program guide and the display format of the night program guide.

  The present invention can be used as an information providing apparatus or the like, for example, as a Web site on the Internet, particularly as an information providing apparatus or the like that distributes information related to content that the user is interested in to the user's terminal device.

It is a block diagram which shows the structural example of the information provision system in Embodiment 1 of this invention. It is a figure which shows the example of the content description information in case a terminal is PC. It is a figure which shows the example of the content description information in case a terminal is a mobile phone. It is a figure which shows the example of user interest information. It is a figure which shows the example of a selection database. It is a figure which shows the example of the log | history information of the user input from a terminal. It is a flowchart which shows the operation | movement procedure of the information provision system until recommendation content information is shown to a user. It is a figure which shows the communication sequence in the information provision system until recommendation content information is shown to a user. It is a flowchart which shows the operation | movement procedure of the information provision system until it learns user interest information from content description information and a user's history. It is a figure which shows the communication sequence in the information provision system until it learns user interest information from content description information and a user's history. It is a figure explaining operation | movement when a user interest information management apparatus and a content recommendation apparatus are arrange | positioned on a home network. It is a figure explaining operation | movement in case the provision apparatus of a content description information provider mediates between a terminal and a content recommendation apparatus. It is a figure which shows the example of the selection database in which the content description information for recommendation and learning became a common item. It is a figure which shows the example of the selection database provided with user ID. It is a figure which shows the example of the selection database provided with the user ID and the content explanation information for recommendation and for learning. It is a figure which shows the example of the content description information produced by the sum of the content description information shown in FIG. 2, and the content description information shown in FIG. It is a figure which shows the example of the content description information produced by the product of the content description information shown in FIG. 2, and the content description information shown in FIG. It is a block diagram which shows the structural example of the content recommendation apparatus of the information provision system in Embodiment 2 of this invention. It is a figure which shows the example of a switching selection database. It is a flowchart which shows the operation | movement procedure of the information provision system until recommendation content information is shown to a user. It is a flowchart which shows the operation | movement procedure of the information provision system until it learns user interest information from content description information and a user's history. It is a figure explaining the operation | movement which selects the content description information suitable for each user from a user's operation history. It is a block diagram which shows the structural example of the content recommendation apparatus of the information provision system in Embodiment 3 of this invention. It is a figure which shows the example of a recommendation system selection database. It is a flowchart which shows the operation | movement procedure of the information provision system until recommendation content information is shown to a user. It is a flowchart which shows the operation | movement procedure of the information provision system until it learns user interest information from content description information and a user's history. It is a block diagram which shows the structural example of the information provision system in Embodiment 4 of this invention. It is a figure which shows the example of program information. It is a figure which shows the example of user interest information. It is a figure which shows the example of the program recommended based on user interest information. It is a figure which shows the example of a list of recommendation programs. It is a figure which shows the process in which the program information to recommend is determined. It is a flowchart which shows the operation | movement procedure of an information provision system. FIG. 10 is a block diagram illustrating a configuration example of an information providing system in a modification of the fourth embodiment. It is a figure which shows the example of the program recommended using the 1st program information. It is a figure which shows the example of the program recommended using program information 1 and 2, and the example of a recorded program reproduction rate. 10 is a flowchart illustrating an operation procedure of an information providing system in a modification of the fourth embodiment. FIG. 20 is a block diagram illustrating a configuration example of an information providing system in another modification of the fourth embodiment. It is a figure which shows the process in which a recommendation program is determined with recommendation priority. It is a block diagram which shows the structural example of the information provision system in Embodiment 5 of this invention. It is a figure which shows the determination process of a recommendation program. It is a figure which shows the example of the operation history with respect to the recommended program. It is a figure which shows the example of a user's operation for every program information. It is a figure which shows the example of calculation of a user's browsing recording rate. It is a figure which shows the process in which a recommendation program is determined based on recommendation conditions. It is a flowchart which shows the operation | movement procedure of an information provision system. It is a block diagram which shows the structural example of the information provision system in Embodiment 6 of this invention. It is a figure which shows the example which synthesize | combines two program information and produces | generates new program information. It is a figure which shows the example which determines a recommended program using the synthetic | combination program information. It is a flowchart which shows the operation | movement procedure of an information provision system. It is a block diagram which shows the structural example of the information provision system in Embodiment 7 of this invention. It is a figure which shows the system configuration example by the apparatus connected with the network. It is a figure which shows the example of the operation history regarding the program content in each apparatus. It is a figure which shows the example of the number of the program information relevant to the definition of the operation history for learning, and the operation history. It is a figure which shows the example of a recommendation program. It is a flowchart which shows the operation | movement procedure of an information provision system. 10 is a flowchart showing an operation procedure of the information providing system in a modification of the first embodiment. It is a figure which shows the example of apparatus specification information. It is a figure which shows the example of a program schedule display specification. It is a figure which shows operation | movement of a program schedule browsing intention estimation part. It is a figure which shows the example of a display of a recommendation program. It is a flowchart which shows the operation | movement procedure of an information provision system.

Explanation of symbols

1, 2, 3 Information providing system 10 Communication network 11 Content explanation information management device 12 User interest information management device 13, 13a, 13b Content recommendation device 14 Terminal 111, 112, 113 Content explanation information 121 User interest information 131, 131a, 131b Content information selection unit 132, 132a Content information recommendation unit 133 Recommended content provision unit 134, 134a Content information learning unit 135 User history information input unit 1310 Selection database 1311 Switching KW content explanation information acquisition unit 1312, 1312a Information switching KW determination unit 1313 Recommended content explanation information acquisition unit 1314 Learning content description information acquisition unit 1315 Provision content description information acquisition unit 1316 Switching selection database 1317 Recommendation method Selection database 3101, 3901, 4501, 5201, 5601 Program information storage unit 3102, 3902, 4502, 5202 Program information storage unit 3103, 3903, 5204 Program information selection unit 3104, 3904, 5205, 5602 User operation history information input unit 3105, 3905, 4504, 5206, 5603 Content information learning unit 3106, 3906, 4505, 5207, 5604 User interest information storage unit 3107, 3907, 4506, 5208 Program content recommendation unit 3108, 3908, 5209, 5606 Recommended program content provision unit 3109, 3909, 4508, 5210, 5607 Program content display unit 3110, 3910, 4509, 5211, 5608 Program content display control unit 3111, 5212 Program operation history storage unit 3911 Program recording operation history storage unit 4303 Program recommendation service 4304 Program recommendation service 4305, 5605 Program content information recommendation unit 4306 Recommended program content control unit 4503 User history information input unit 4507 Recommended program content control provision unit 4510, 5610 Recommended program operation history storage unit 4511 Recommended condition determination unit 5203 Program information composition unit 5609, 5701 User operation history storage unit 5611 Learning operation history selection unit 5702 User operation input unit 5703 Device control unit 5704 Content display unit 5706 Content storage unit 6201 Device Specification information storage unit 6202 Program guide display specification storage unit 6203 Program guide browsing intention estimation unit

Claims (16)

An information providing device that provides information related to content that a user is interested in to the terminal device of the user,
Accepting means for accepting access by the terminal device of the user via a communication network;
Content description information management means for holding a plurality of types of content description information describing a description of a plurality of types of content associated with each of a plurality of different types of terminal devices ;
User interest information management means for holding user interest information which is information relating to the interests of a plurality of users;
Recommended content description information used for determining content of interest to the user from among a plurality of types of content description information held by the content description information management means based on the type of the terminal device of the user who has accessed And content explanation information selection means for selecting and reading provision content explanation information for providing to the user ,
By referring to the user interest information held in the user interest information management means, information relating to the interest of the accessing user is identified, and the recommended content description read by the identified information and the content explanation information selection means based on the information, the recommended content determining means for determining a content to be recommended to the user,
Providing means for identifying information relating to the determined recommended content from the providing content explanation information read by the content explanation information selecting means, and providing the identified information to the terminal device of the user. Information providing device. The content description information selection means has a selection database that is a table in which the type of terminal device is associated with recommendation content description information and provision content description information, and the recommendation database is referred to by referring to the selection database. information providing apparatus according to claim 1, wherein the selecting the content description information and the provided content description information. In the selection database, for each of a plurality of users, the type of terminal device, recommendation content description information, and provision content description information are associated with each other,
The content description information selection means refers to the selection database, and determines the recommended content description information and the provision content description information corresponding to the combination of the user who has accessed and the type of the terminal device of the user. The information providing apparatus according to claim 2 , wherein the information providing apparatus is selected. The user interest information includes an interest keyword indicating an interesting term for each of the plurality of users.
The recommendation content explanation information includes an explanation keyword indicating terms related to the explanation of the content for each of the plurality of contents.
The recommended content determining means, accessed by matching the interest keywords and the description keyword corresponding to the user has, the information providing apparatus according to claim 1, wherein determining the content. The information providing apparatus further includes:
User history information input means for acquiring history information related to content selection by the user;
Based on the acquired history information, the information providing apparatus according to claim 1, characterized in that it comprises a learning means for updating the user interest information stored in the user interest information management unit. The content description information selection means is further for learning used for updating the user interest information from among a plurality of content description information stored by the content description information management means based on the type of the terminal device of the accessing user. Select and read content description information,
The information providing apparatus according to claim 5 , wherein the learning unit updates the user interest information based on the content description information read by the content description information selection unit and the history information. The user interest information includes, for each of the plurality of users, an interest keyword indicating a term of interest and weight information regarding its weight,
The history information includes weight information indicating the degree of interest of the user in content selection,
The recommended content determination means determines the content based on an interest keyword and weight information corresponding to a user who has accessed,
The information providing apparatus according to claim 5 , wherein the learning unit updates weight information of the interest information based on weight information included in the history information. The content description information management means includes a content description information creation unit that creates new content description information from the plurality of content description information,
The content description information selection means selects the recommended content description information and the provision content description information from a plurality of content description information including the content description information created by the content description information creation unit. information providing apparatus according to claim 1, wherein. The information providing apparatus according to claim 8 , wherein the content description information creating unit creates the content description information by calculating a sum or product of information included in the plurality of content description information. The content description information selection means includes
A switching content description information acquisition unit for acquiring switching content description information used for switching content description information from among a plurality of content description information held in the content description information management means;
A switching determination unit that determines whether or not to switch the recommended content description information, the providing content description information, and the learning content description information to be selected based on terms included in the acquired switching content description information;
Claim 6 when it is determined that the switch, a new recommendation content description information, and having an after-switching content description information acquisition unit for acquiring and selecting the provided content description information and learning contents described information The information providing apparatus according to the description. The switching content description information includes a keyword indicating a term describing the content,
The content description information selection means further holds a switching selection database indicating correspondence between the switching keyword and the recommended content description information after switching, the providing content description information, and the learning content description information for a plurality of users. ,
The switching determination unit, by collating the switching keyword included in the keyword and the switching selection database included in said switching content description information, information according to claim 1 0, wherein that said determining Providing device. The recommended content determination means determines the content according to any of a plurality of predetermined recommendation methods,
The content description information selection means includes
Switching for acquiring switching content description information used for switching a recommended method used for determining recommended content by the recommended content determination unit from among a plurality of content description information held in the content description information management unit Content description information acquisition unit,
A switching determination unit that determines whether or not to switch the recommendation method based on terms included in the acquired switching content description information;
The recommended content determining means, wherein when it is determined that the switching by the switching determination unit, according to the new recommendation mode after switching, the information providing apparatus according to claim 1, wherein determining the content. The learning means updates the user interest information according to any of a plurality of predetermined recommendation methods,
The content description information selection means includes
For switching content description information for switching used for switching a recommendation method used for updating the user interest information by the learning unit from among a plurality of content description information held in the content description information management unit A content description information acquisition unit;
A switching determination unit that determines whether or not to switch the recommendation method based on terms included in the acquired switching content description information;
The information providing apparatus according to claim 6 , wherein the learning unit updates the user interest information according to a new recommendation method after switching when the switching determining unit determines to switch. The switching content description information includes a keyword indicating a term describing the content,
The content description information selection means further holds a recommendation method selection database indicating a correspondence between a switching keyword and a recommended method after switching for a plurality of users,
The switching determination unit, by collating the switching keyword included in the keyword and the switching selection database included in said switching content description information, according to claim 1 2 or claim, characterized by the determination 13. The information providing device according to 3 . An information providing method by an information providing device for providing information related to content that a user is interested in to the terminal device of the user,
A receiving step in which the receiving means included in the information providing device receives access by the terminal device of the user via a communication network;
Content explanation information management means for holding a plurality of types of content explanation information in which the content explanation information selection means provided in the information providing device describes explanations concerning a plurality of types of content associated with a plurality of different types of terminal devices. Based on the type of the terminal device of the accessing user, the recommended content description information used for determining the content that the user is interested in and the providing content description information for providing to the user are selected. Content description information selection step to be read out,
The recommended content determination means provided in the information providing device identifies information related to the interest of the user who has accessed by referring to the user interest information management means that holds user interest information that is information related to the interest of a plurality of users , based on the recommended content description information read by the specified information and the content description information selection step, the recommended content determining step of determining a content to be recommended to the user,
The providing means included in the information providing device identifies information regarding the determined recommended content from the providing content description information read out by the content description information selection step, and the identified information is stored in the terminal device of the user. An information providing method comprising: providing a providing step. A program for an information providing device that provides information related to content that a user is interested in to the terminal device of the user,
A program causing a computer to execute the steps described in the information providing method according to claim 15 .

Images