JP2007286817A - Information delivery system, information delivery device and information delivery method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform filtering of information based on a user's ordinary action on the Internet. <P>SOLUTION: An information delivery server comprises an integrated feed generation means for performing redirect processing for performing connection to each WEB site or the like which provides feeds via the information delivery server to each feed to be delivered to a user terminal, and generating an integrated feed composed of a plurality of feeds subjected to redirect processing; a browsing control means for storing, upon receipt of a browsing request to each feed of the integrated feed from the user terminal, the browsing request to the feed concerned as the user's browsing history in a browsing history database, and a means for generating user network information by mutually associating a plurality of users having the similar browsing history for each feed. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information distribution system that distributes information suitable for a user's preference. More specifically, the present invention acquires an action history (browsing history) on a user's network without imposing a burden on the user, and The present invention relates to an information distribution system, an information distribution apparatus, and an information distribution method for performing information distribution suitable for user preferences based on the information distribution.

  Usually, humans naturally sort information that is necessary for them, information that is necessary for friends and friends, and information that is not so in the network of human relationships, and humans themselves create order and classification for information. .

  However, on the Internet (WEB), countless information is flooded in a form that is almost chaotic, and it is for the user that the user selects the countless information, selects the information of interest, and collects the desired information. The current situation is a big burden. For example, even if a keyword search is performed using a search engine, the preference, reliability, and utility of the hit information for the user are not ensured. That is, since effective information filtering cannot be performed, the user cannot easily reach useful information.

  On the other hand, from the viewpoint of providing information, even if the user provides useful and reliable information to the user, if the user cannot reach the information, the value of the information itself is reduced. Information transferability is impaired.

  On the other hand, the reliability and usefulness of information is determined by the user himself / herself, other users who have the same interests / preferences, acquaintances, friends, or persons or specialists who are knowledgeable in a specific field It depends greatly on whether or not the information is recommended by a person who is recognized by the surroundings. That is, a person who has the same interest / preference, a person whom the user himself trusts, a person who recognizes the user, gives the user a sense of trust or empathy. Therefore, information recommended by these people can be filtered information that is already reliable and highly useful for the user. In addition, if the information acquired in this way is useful to the user, the opportunity to recommend and introduce it to other people, acquaintances, friends, etc. will increase, and the information will be more people. Will lead to propagation.

  There is an SNS (Social Networking Site) type network community that exchanges information by connecting the same interests and preferences on the Internet. This SNS-type network community connects each user on the Internet as a friend and acquaintance, and exchanges messages, exchanges information, and acquaintances on the Internet in such a way that each user's personal relationship network can be seen on the Internet. Various actions such as introduction and recommendation can be performed, and information and news that the network user (member) is paying attention to can be found through the user's diary and blog (Weblog) in the network. That is, in this SNS type network community, information exchange in a network constructed by friends, acquaintances, and people with the same interests and preferences is realized. Information filtering functions such as order creation and classification will be provided to users. A typical SNS network community site is the social networking service “mixi” (registered trademark, http://mixi.jp/).

  In addition, from the viewpoint of the usefulness of information, there is also an information evaluation and information introduction system by word of mouth. For example, by distributing information (evaluation information) posted by a large number of unspecified users on specific products and information, Providing useful information to users while improving propagation.

  Further, collaborative filtering has been conventionally used as an information filtering technique. This collaborative filtering technology extracts the user's preference from the user's past behavior history on the Internet, and guesses the user's preference based on the preference information of the user who is acting similar to the user In addition, the filtering technique provides only preferable information of preference. For example, in the latest article recommendation method described in Patent Document 1, this collaborative filtering technique is utilized, and a plurality of similar users are extracted from user history information, and subcategory candidates to be recommended are extracted by analyzing them. Provides the latest articles based on this subcategory candidate.

  Furthermore, in recent years, there is a feed technology for easily acquiring information watch and update information on WEB. A feed is information attached to content provided from an XML format WEB site based on standard specifications such as RSS, RDF, and Atom (information source such as URL of the WEB site, attribute of provided information (WEB site) The date and time, title, classification, etc.) of the information and articles to be disclosed)), and is usually obtained using a feed reader such as an RSS reader. The feed reader has an automatic patrol function, and once a user registers a feed to be subscribed to, the feed registered by this automatic patrol function is automatically acquired, so the site you want to read automatically It is possible to obtain the update information (or obtain a notification of the update). This feed is not the content itself, but XML format data that provides the user with the location of the content (information source) and the attributes of the information (date, title, classification, etc.). The type of information such as video information) is not questioned.

Japanese Patent Laid-Open No. 2003-216636 (FIG. 1, FIG. 3, etc.)

  However, in the SNS type network community, information senders who send information must post (register) information actively in a diary or blog, and similarly, information introduction by word of mouth also introduces information. The information introducer's active action is necessary, such as the information introducer sending an email or a message.

  In addition, participation in the SNS network community requires actions such as recommendation and introduction from users belonging to the network. That is, users are not related by information but are related by the new density and reliability of users, so the SNS type network community has information that is exchanged in the network is closed. It becomes a strong tendency to become information exchange with limited information, and the user network cannot have a sufficient spread.

  Further, in the information recommendation system using the conventional collaborative filtering, the function of acquiring the user's action history information on the Internet and the function of displaying the recommendation result of the information based on the action history information are specified by a specific user or plug. In (program etc.), it depends on the information media, and it is necessary to generate HTML that highlights the result of recommending information by the program and to have the user visit the site. That is, since users, information media, and browsing sites are limited, the number of users and the number of contents do not increase so much in conventional collaborative filtering. For this reason, there is a problem that it is not possible to sufficiently provide information to a user by giving order to a large amount of information and performing narrowing down.

  In collaborative filtering, the result of recommending information is not provided in a form that is easy for the user to use. Therefore, the user has to bother to see the recommendation result using a WEB browser or client software. That is, even if notification that there is information recommended to the user by e-mail or the like, in order to view the information selected by collaborative filtering, it is necessary to access a user-specific page on the WEB, etc. Therefore, although it is not necessary to browse each WEB site individually with a WEB browser, the burden on the user is somewhat reduced, but the burden on the user is not reduced much. It is necessary to distribute information that minimizes the burden on the user and minimizes unnecessary actions such as access to the user's information.

  Furthermore, since the feed technology is data in XML format that carries the location and attributes of the content instead of the content itself as described above, the information medium of the content itself is not questioned. For this reason, the information management by feed reduces the burden on the user compared with the case where the user visits and browses the WEB site registered in the bookmark of the browser one by one, but when the number of feeds increases, the acquisition registered by the user As a result, there is an increase in the number of feeds related to the information that the user wants to do. Eventually, the user cannot organize the information. ing.

  The present invention has been made for such a problem, and the purpose of the present invention is to perform an action on the Internet that a user routinely performs, that is, information selection based on information selection or information subscription actions. An object of the present invention is to provide an information distribution system, an information distribution apparatus, and an information distribution method capable of performing filtering, reducing the burden on the user, and distributing only useful information to the user.

  In particular, in the present invention, a user network that is not limited by the user attributes and information format of the user who browses information is constructed, and an information distribution system capable of providing a wide range of information suitable for the user is realized. Objective.

  An information distribution system as one aspect of the present invention is an information distribution system including an information distribution server that collects feeds provided from a WEB site on the Internet and distributes the feeds to a user terminal of a user, An information distribution server distributes a plurality of feeds based on a plurality of feed distribution registration information registered via a user terminal from a web site on the Internet providing each feed, and the user terminal. Aggregated feed obtained by performing redirection processing for connecting to each WEB site that provides a feed via the information distribution server and aggregating a plurality of feeds subjected to the redirection processing The aggregate feed control means for generating the When the browsing request for the feed is received, the browsing control means for storing the browsing request for the feed in the browsing history database as the browsing history of the user, and user network information that associates a plurality of users with similar browsing history for the feed is generated. User network construction means, and recommended feed control means for generating a recommended feed for recommending a feed to the user based on the browsing history of each user associated with the user network information.

  According to the present invention, a plurality of feeds registered by the user are aggregated into one and distributed to the user terminal. Therefore, even if the user does not acquire feeds individually from each feed, the user collects a plurality of feeds at once. Since each feed that can be acquired and distributed to the user is subjected to redirect processing for connecting to each WEB site that provides the feed via the information distribution server of the present invention. Even without any special mechanism or user operation, it is possible to acquire browsing history (information selection, information browsing request, etc.) on the Internet that the user performs on a daily basis.

  And since a user network with a plurality of other users is constructed in association with the feed viewed by each user, it is beneficial to each user without being limited to the user attributes and network attributes. It is possible to distribute and recommend high information.

  That is, the information distribution system of the present invention can provide the user with information selection, order, classification, and the like suitable for the user's preference, and can collect and distribute information useful to the user. .

  Hereinafter, an information distribution system realized by an information distribution server to which an information distribution program (information distribution method) of the present invention is applied will be described in detail according to the following preferred embodiments with reference to the accompanying drawings. .

  FIG. 1 is a schematic configuration diagram of an information distribution system in an embodiment of the present invention. The information distribution system of the present embodiment includes an information distribution server 100 that collects feeds provided from a plurality of WEB servers (WEB sites) 300 on the Internet and distributes the collected feeds to the user terminals 200 of the respective users. Yes. The user terminal 200 is a computer having communication means connectable to the Internet, and may be a mobile terminal or a mobile phone. The user terminal 200 is installed with a feed reader such as an RSS reader for receiving a distributed feed and browsing or subscribing to the feed, and a WEB browser for browsing a WEB page provided by the WEB server 300. Has been. The WEB server 300 is a content providing server that distributes content such as information, articles, music, and videos. In order to widely distribute these information to users, a feed is generated for each content.

  Normally, a feed reader such as that described above is used to view a feed. This feed reader is equipped with a feed registration function, a feed automatic patrol collection function, and a feed display function, and provides a feed (feed provided) desired to be browsed. URL information of the WEB server to be registered) in this feed reader, when the feed reader is activated in the user terminal 200, the latest information is sent from each WEB server 300 that provides the feed registered automatically. The feed is collected, and the information included in the feed or the summary of the article and the article title are displayed on the display unit of the feed reader. Therefore, the user can browse the latest feed from the user terminal 200, and by selecting (clicking etc.) an article (article title) of this feed with an operation means such as a mouse, based on the URL information included in this feed. A WEB server (WEB site) 300 that provides the feed can be accessed.

  The feed is information accompanying the content provided from the XML format WEB site based on the standard specifications such as RSS, RDF, and Atom as described above, and the WEB site URL information (information source information), It includes information and article attributes (information and article dates, title information or summary information, classification related to information and articles (tag information)) provided by the WEB site. For example, if an economic feed is provided from a WEB site, this feed will include the latest information or article title information or summary information, and “finance, stocks” to correlate this feed with economic information. Tag information such as “market price, yen market price, bank, interest rate” (keyword information indicating the category of information of the feed and the type of information, and a plurality of feeds are associated with one tag information). And URL information of the feed provider (information sender). The feed is XML format information in which a plurality of articles are associated with the feed header of the WEB site that provides the feed. For example, the feed A includes a plurality of articles 1, articles 2,. Each article is configured as an XML-format article item (article item) composed of the date and time when the article was posted, the title or summary of the article, URL information, and tag information. Therefore, browsing the feed includes browsing a plurality of article items included in the feed, and the processing for the feed includes processing for each article item as its concept.

  The information distribution system of the present embodiment reduces the burden on the user and distributes useful information to the user while utilizing such an existing feed reader, and the user attributes and information format of the user who browses the information A user network that is not limited to the above is constructed to provide a wide range of information suitable for the user.

  FIG. 2 is a configuration block diagram of the information distribution server 100, a feed distribution unit 101 that controls feed distribution to the user terminal 200, feed registration information registered by the user via the user terminal 200, and tag information included in the feed In response to a feed collection request from the information distribution database (viewing history database) 102 and the user terminal 200 in which the user's feed browsing history and user network information are stored, each user's feed stored in the information distribution database 102 A feed collection unit 103 that collects the latest feed from each WEB server 300 based on the registration information, a feed control unit 104 that receives a feed browsing request from the user terminal 200 and stores a user browsing history in the information distribution database 102. , For each feed, for each user A feed importance calculation unit 105 that calculates a feed importance that serves as a reference for determining whether the information is important (beneficial) information, and a collaborative filtering function are provided. A user network construction unit 106 that generates or constructs user network information that associates users with a high degree of similarity, and a feed that the user desires to distribute, and performs various settings in feed acquisition A registration control unit 107 that provides the user terminal 200 with a registration screen (WEB page) and stores the registration information and setting information input via the WEB page in the information distribution database 102 as feed distribution registration information of each user. It has. These units are controlled by a control unit (CPU) 108.

  Next, each configuration of the information distribution server 100 of this embodiment and process transition of the information distribution system will be described.

  First, as shown in FIG. 3A, the information distribution server 100 according to the present embodiment is a URL for feed distribution to users, that is, a feed distribution registration for receiving distribution of a desired feed to each user. A WEB page is provided (step S301). The user registers the feed distribution registration URL in the feed reader of the user terminal 200 (step S101). The user can refer to the WEB page provided by the information distribution server 100 based on the feed distribution registration URL via the WEB browser (or via a feed reader equipped with a WEB browser function) (step S102). ).

  FIG. 7A is an example of a WEB page (user-dedicated WEB page) provided via the WEB browser of the user terminal 200. The user selects and registers a feed for which feed distribution is desired from a list of feeds that have been hit by feed search (keyword search or tag search), or the URL of the WEB server (WEB site) that provides the feed by the user in advance. If the URL is known, the feed desired to be distributed is registered in the information providing server 100 by directly inputting the URL from the input means (step S103). The feed distribution registration URL and the WEB page are provided by the registration control unit 107, and information selected or input via the WEB page is used as the feed distribution registration information of each user by the registration control unit 107. 102 (step S302).

  The feed distribution unit 101 includes an aggregate feed control unit 101a and a recommended feed control unit 101b. The feed distribution unit 101 is configured to execute an aggregate feed request by a user operation from the user terminal 200 or when a feed reader is activated in the user terminal 200. The aggregate feed request transmitted automatically is received (steps S104 and S303, see FIG. 3B), and the type of distribution feed requested by the user is determined (step S304). The determination processing in step S304 is based on the request type (identification information such as a flag indicating whether an aggregate feed is requested or a recommended feed is requested) transmitted from the user terminal 200 as a feed request. Done. When the feed distribution unit 101 determines in step S304 that the user has requested an aggregate feed, the feed distribution unit 101 outputs an instruction to request the aggregate feed processing to the aggregate feed control unit 101a, and the aggregate feed control unit The process proceeds to the aggregate feed generation process by the 101a and the feed collection unit 103.

  In response to the processing request from the feed distribution unit 101, the aggregate feed control unit 101a acquires the feed distribution registration information of the user who has requested the feed from the information distribution database 102, and each feed included in the feed distribution registration information The feed collection unit 103 is requested to collect the feed based on the URL (step S305). The feed collection unit 103 performs a connection process to the WEB server 200 that provides the feed based on the URL of each feed, and acquires the latest feed provided by the WEB server 200 (step S307). Then, the aggregate feed control unit 101a performs redirect processing on each feed (URL for each of a plurality of article items included in the feed) collected by the feed collection unit 103 (step S308). Specifically, if the URL 1 of the information distribution server 100 of the present embodiment is “www.xxx.co.jp”, the URL 1 of the information distribution server 100 is added to the URL 2 of the collected feed, for example, If the URL 2 of the feed A is “www. △△△△ .co.jp”, the URL 3 of the feed after the redirect processing is “www. ○○○○ .co.jp / www. △△△△. co.jp ". Furthermore, in the present embodiment, this redirect processing process adds a parameter such as an identification ID of a feed to be distributed to this URL 3 in consideration of a user browsing history acquisition process. For example, a feed identification ID is assigned in advance to each feed of the feed distribution registration information stored in the information distribution database 102, and URL4 = “www.xxx.co.jp/ Feed identification ID / www.ΔΔΔΔ.co.jp ”is generated. In addition to the feed identification ID, an ID of a user to whom the feed is distributed or an ID for each article item can be applied as a parameter. Details will be described later.

  Thereafter, the aggregated feed control unit 101a performs a process of combining and aggregating the feeds subjected to the redirect process, and generates one aggregated feed in the XML format (step S309). Specifically, since the XML-format feed provided from each WEB site is configured as feed header + article 1 + article 2... Of each WEB site, the aggregate feed generation unit 101a of the present embodiment Extract only the articles for each feed with the feed header removed for each feed. Each of these articles formatted in the XML format as described above constitutes one article item including the date and time when the article was published, the title or summary of the article, URL information, and tag information. Each article item from which the feed header is removed is extracted from each feed, the extracted article items are aggregated into one, and an aggregate feed header corresponding to each user is added to generate an aggregate feed. The aggregated feed generated in this way is delivered to the user terminal 200 by the feed delivery unit 101 (step S310). That is, in the information distribution system of the present embodiment, the information distribution server 100 does not acquire each feed individually from the WEB site that provides the feed, but the individual information distribution server 100 registers the feed. A plurality of feeds based on the feed distribution registration information collected are collectively collected, and the collected feeds are aggregated and provided to the user terminal 200 (see FIGS. 1 and 6).

  When the feed distribution unit 101 determines in step S304 in FIG. 3B that the user has requested a recommended feed, the feed distribution unit 101 instructs the recommended feed control unit 101b to request a recommended feed process. , And transition to recommended feed generation processing by the recommended feed control unit 101b, the feed collection unit 103, and the feed importance calculation unit 105.

  In response to the processing request from the feed distribution unit 101, the recommended feed control unit 101b makes a recommendation to the user based on a user browsing history, user network information, feed importance, and user feed distribution registration information, which will be described later. A recommended feed is extracted, and the processing from step 307 to step 309 in FIG. 3B based on the extracted recommended feed is performed, and the recommended feed generated by the feed distribution unit 101 is distributed to the user terminal 200 (step S310).

  FIG. 8 is a diagram illustrating an example of a feed reader of the user terminal 200. The above-described feed distribution registration URL is registered in the feed reader, and an aggregate feed ((Aggregation)) is displayed in the feed list display unit FD of the feed reader. One recommendation feed and one recommendation feed (Recommendation) are displayed. The feed request in step 104 of FIG. 3B is performed by selecting an aggregate feed or recommended feed displayed in the feed list, so that the feed request of the information distribution server 100 is transmitted, and information distribution from step S303 to step S310 is performed. Through each process of the server 100, the aggregate feed and the recommended feed are distributed to the user terminal 200.

  Further, the feed reader of FIG. 8 is provided with a title list display unit TD for displaying a plurality of information included in the aggregated feed or recommendation feed and title information (article item) of the article, and a WEB browser unit WD. When the user selects the title information of the title list display unit TD with a selection means such as a mouse, a feed browsing request regarding the selected article item is transmitted to the information distribution server 100 (see step S105, FIG. 3C), and WEB A WEB page from the WEB server 200 that distributes the selected feed is displayed on the browser unit WD.

  The browsing control unit (browsing request receiving unit) 104 of the information distribution server 100 receives a browsing request (feed browsing request) for each article item of the aggregate feed or recommended feed displayed on the feed reader. I do. Specifically, the redirect response process is performed in response to the reception of the feed browsing request (step S311), and the browsing history of the article item browsed by the user (feed browsing history) is used as the user browsing history in the information distribution database. 102 (step S312). Further, the user network construction unit 106 is instructed to execute a user network construction (user network information generation and update) process based on the user's feed browsing request (browsing history) (step S313).

  As shown in FIG. 1, the acquisition processing of the user browsing history in the information distribution system according to the present embodiment is performed by redirect processing on each article item of the aggregate feed. The browsing request for each article item is not transmitted to the WEB site 300 that provides the feed to which the article item that has been directly browsed (route C), but is transmitted to the information distribution server 100 of the present embodiment ( Route A). Then, the information distribution server 100 performs a redirect response process to display a WEB page that provides the feed on the WEB browser of the user terminal 200 (route B), and the browsing history storage process for the feed by the browsing control unit 104 and the user The network construction unit 106 generates (updates) user network information.

  Therefore, in the present embodiment, the WEB server 200 is accessed via the connection route A and route B, but the user side seems to be directly connected to the corresponding WEB server 200 based on the selected feed browsing request. WEB site browsing. That is, in the information distribution system according to the present embodiment, the user can browse the feed by selecting a daily feed when the user browses the information, connecting to the WEB site, or the like without requiring any special operation or application. The history can be acquired.

  Next, the browsing request process and the recommended feed generation process of the information distribution system of the present embodiment described above will be described in detail with reference to FIGS.

<Browsing request processing>
As shown in FIG. 4, the browsing request process of the information distribution system according to the present embodiment includes a redirect response process and a browsing history storage process performed by the browsing control unit 104 after receiving a feed browsing request from the user terminal 200. The user network construction unit 106 includes two processes of generating and updating user network information. As described above, the aggregate feed and the recommended feed (each article item included in the aggregate feed and the recommended feed) distributed to the user terminal 200 are subjected to redirect processing, so that the feed reader of the user terminal 200 When the displayed feed is selected by the user, the WEB server is not indirectly connected to the WEB server 200 that directly provides the feed (content) but indirectly through the information distribution server 100 of the present embodiment. 200. Therefore, the user's browsing history for the feed such as which feed the user has selected can be acquired. In particular, since the feed parameters (article items) distributed to the user as described above are provided with browsing parameters such as each feed identification ID, the ID of the user to whom the feed is distributed, and the ID of each article item. The browsing control unit 104 acquires the browsing parameters included in the URL when there is a connection request from the user terminal 200 based on the feed browsing request from the user, that is, the URL subjected to the redirect processing, The browsing history of the user is stored in the information distribution database 102 based on the browsing parameters.

  The browsing history stored in the information distribution database 102 includes a browsing history for a feed indicating which user has viewed what feed (when date and time) and how many times, and which user belongs to which feed and when (date and time). A browsing history indicating whether information or an article has been browsed and a browsing history indicating how many times a user has browsed a feed with which tag information is included. For example, when user A was viewing feed A twice, when and when, feed A was viewed by user A and user B twice, three times, and when the feed has tag information X Is stored in the information distribution database as a browsing history for the user's feed, such as when the user A and the user B browsed 4 times, 6 times, and when.

  Hereinafter, the browsing request process of the present embodiment will be described in detail with reference to FIG. As shown in FIG. 4, first, when a user selects an article item desired to be browsed displayed on the title list display section TD of the feed reader of the user terminal 200 via the mouse or the like, the WEB browser of the feed reader The connection function to the Internet of the part WD connects to the information distribution server 100 of the present embodiment according to the URL subjected to the redirect processing included in the article item. The browsing control unit 104 receives the access from the user terminal 200 as a browsing request for the user's feed. Upon receiving the user's feed browsing request, the browsing control unit 104 transmits a response for HTTP redirection to the WEB server (WEB page) that provides the feed requested by the user to the WEB browser unit WD of the user terminal 200. (Redirect response processing), this feed browsing request is stored in the information distribution database 102 as user browsing history data.

  The browsing control unit 104 outputs a generation request instruction so as to generate (update) the user network information to the user network construction unit 106 after completion of the processing for storing the browsing history of the user. Upon receiving the generation (update) processing request from the browsing control unit 104, the user network construction unit 106 acquires browsing history data and feed distribution registration information of all users registered from the information distribution database 102.

  Then, the user network construction unit 106 calculates the similarity between the plurality of users based on the acquired browsing history and the feed distribution registration information, that is, the user's preference level based on the browsing history by the collaborative filtering function, The similarity of the user's preference level is calculated. Then, user network information in which users having a predetermined similarity are associated with each other is generated, and this user network information is stored in the information distribution database 102. In the present embodiment, a plurality of users having a high degree of similarity are extracted, and a feed having a high browsing frequency or tag information having a high selection frequency of each user in the user network extracted from the browsing history of these users is separately added. However, for example, all similarities regarding a specific user A and all other users (similarity of user B viewed from user A, similarity of user C, User D's similarity, ... and vice versa) is stored as user relational information, and when using user network information, the user based on this similarity is extracted, and the extracted user's By acquiring each browsing history from the information distribution database 102, user network information can be generated as appropriate without being stored in a separate table or the like. For this reason, the user network information of this embodiment should just be the information containing the similarity of the preference between all the users at least. Note that user network construction (generation and update of user network information) by the network construction unit 106 is performed whenever a browsing history is added or updated, that is, a feed browsing request from a user is received, and the latest in real time. Build a user network.

  As described above, the user network construction of the user network construction unit 106 of the present embodiment is performed by associating a plurality of users having similar user preferences by collaborative filtering based on the browsing history of each user for each feed. . The user's preference is calculated from the common points of the three elements including the user's browsing history. First, to which feed the user is registered to request distribution (feed distribution registration information), second, to which feed the user has browsed, and third, the browsed feed is The user's preference is calculated in consideration of which tag information is used as a common element of the user preference.

  For example, when the feed A is registered in the feed distribution registration information of the user X and there is a browsing request for the feed A, the user's preference for the feed A is high. Thus, by calculating the user's preference level for all the feeds, it is possible to calculate the degree of similarity between users. That is, all users whose preference level for the feed A is within a predetermined value are extracted, and the similarity level between the user and a plurality of other users is calculated between 0 and 1 according to the preference level. Similarly, by extracting the browsing for which tag information is browsed from the browsing history of the user and calculating the user's preference for each tag information, the similarity between users associated with the preference for the tag information can be obtained. Can be calculated. Therefore, the user network information generated by the user network construction unit 106 is associated with a plurality of users having high similarities with each other, and the browsing history of the plurality of users, that is, each user in the user network. The browsed feed and the tag information of the browsed feed are included (the user network information holds the feed and tag information that the user network is interested in).

  In addition to the user network that is automatically constructed without the user's awareness based on the similarity between users using the browsing history acquisition and collaborative filtering functions as described above, However, it is possible to construct a network between users consciously.

  That is, other users that the user himself / herself pays attention through the registration screen provided by the registration control unit 107 of the present embodiment, for example, friends who are already known to have the same hobbies and interests, prominent experts, identification An intentional user network can be constructed by the user intentionally registering a person who is familiar with the field in the information distribution server 100. For example, like the WEB page for feed distribution registration, a user registration screen as shown in FIG. 7B is provided by the registration control unit 107, and an intentional user registered via the user registration screen is displayed in the information distribution database. 102.

  As described above, the user network provided by the information distribution system of the present embodiment is that the user is aware of the intentional user network in which the user intentionally constructs the network and the user's daily action history on the Internet. It is used for recommended feed generation processing by a recommended feed control unit 101b and the like which will be described later.

<Recommended feed generation processing>
The recommended feed generation process of the information distribution system according to the present embodiment is based on the automatic user network constructed by the above browsing request process, the intentional user network constructed by the user consciously, and the browsing history of the user. This is a process for recommending a feed (article item) useful to the user and providing it to the user.

  Hereinafter, the recommended feed generation process of this embodiment will be described in detail with reference to FIG. In response to the processing request from the feed distribution unit 101, the recommended feed control unit 101b obtains the user network information of the user who recommends the feed constructed by the above-described browsing request processing of the present embodiment and the browsing history of the user. Obtained from the information distribution database 102. Then, a recommendation candidate feed is extracted from the acquired user network information and browsing history of the user.

  Specifically, the recommended feed control unit 101b extracts tag information included in the user network information constructed (generated) by the user network construction unit 106 and has been browsed by users in the user network so far. For example, user network information for user A includes user X (similarity 0.5), user Y (similarity 0.3), and user Z (intentional user network registered by user A intentionally). The tag information included in the browsing history that the user A has browsed is “computer”, “gourmet”, and the tag information that the entire user network is interested in, that is, It is assumed that tag information that has been browsed so far by four users, user A, user X, user Y, and user Z, is “computer”, “gourmet”, and “car”. In this case, the tag “car” (a plurality of feeds related to the tag “car”) has never been viewed by the user A.

  Therefore, the recommended feed control unit 101b extracts, as recommended candidate feeds, tag information that is interesting to the user network of the user who makes the recommendation and is not in the user's feed distribution registration information or browsing history. Further, in the present embodiment, the tag information that the user network of the recommending user is interested in and the feed that is not in the user's feed distribution registration information or browsing history is distributed to any user other than the user network. Even in such a case, the feed is automatically extracted as a recommended candidate feed.

  As described above, the extraction of the recommended feed for the user who makes the recommendation according to the present embodiment uses the tag information, so that it is possible to widen the feed that is a candidate for the recommendation candidate. Among feeds including tag information of “Computer” that is interested in the user network, feeds related to “Computers” that have not been viewed by User A (User A does not know) (or articles included in the feed) When the item) is attracting attention by other users than the user network, the user A is likely to want to browse and subscribe to this feed, which is highly useful information for the user A. Conversely, feed information and articles that do not have common points by such tag information are unlikely to be read from the beginning by the user A.

  Then, the recommended feed control unit 101b outputs a feed collection request to the feed collection unit 103 and acquires a recommendation to the feed importance calculation unit 105 in order to obtain the latest feed of the extracted recommendation candidate feeds. A feed importance calculation request for calculating the importance of the extracted recommended candidate feed for the user who performs the operation is output.

  As described above, the feed collection unit 103 acquires the URL of the feed to be collected from the information distribution database 102 and acquires the feed from the WEB server 200 that provides the feed.

  On the other hand, the feed importance level calculation unit 105 calculates the feed importance level indicating how much the feed that has become a recommendation candidate for the user who makes the recommendation is useful for the user and that the user desires to subscribe. This feed importance is calculated for each of the feeds extracted as recommendation candidates, and is calculated based on the browsing history of each user's recommendation candidate feed in the user network. Specifically, firstly, in the past overall browsing history for the feed, the percentage of browsing by each user of the user network, and secondly, in the browsing history of the recommended candidate feed in the user network, A browsing history of the user A who has a higher degree of similarity (user X than user Y) or thirdly, the recommended candidate feed is browsed by user A himself and user Z of user A's intentional user network. The importance for the user A with respect to each recommended candidate feed is calculated. In particular, the more time it has passed since each user last read their recommendation candidate feed, the less useful the information is, so how many users in the user network have subscribed and recently subscribed It is preferable to consider whether it is a thing as an element of calculating the importance. This feed importance level is calculated by Formula I shown in FIG. 9, and the number of browsing times (v (n, f)) with respect to the recommended candidate feed, the number of browsing times of users with high similarity in the browsing frequency (similarity level (s (X, n))), whether the date and time of browsing is the latest (t (n, f)) is an index of importance for each feed for the user. Also, as described above, as the time has elapsed since the last time the recommended candidate feed was read, the usefulness of the information is impaired. Therefore, whether the date and time of browsing is the latest is Exp (t0-t (N, f)) (the current time is t0, and the time when the user last viewed the recommended feed is t (n, f)). Also, the parameter A in Formula I in FIG. 9 is an index indicating how important the user network intentionally constructed by the user is, and by raising this parameter A, the user A of the intentional user network The importance of a feed in which a recommendation candidate feed is viewed is increased. Usually, since the user network is intentionally constructed by the user, it is preferable to set this parameter A to 1 or more. In Formula I of FIG. 9, the denominator has the total number of past browsing for the feed (v (f)). This is simply the importance of the calculated feed (each article item) to the user. Is calculated as a ratio, and it is arbitrary whether v (f) is applied to the denominator of Formula I.

  Next, the recommended feed control unit 101b finally selects an article item to be recommended to the user based on each feed importance level of the recommendation candidate feed calculated by the feed importance level calculation unit 105. A feed having a high feed importance may be distributed to the user as it is, but in this embodiment, the feed is further selected using the number of browsing for each article item of the recommended candidate feed. In other words, the number of browsing for each article item of the recommended candidate feed (for example, the number of browsing for each article item within one week) is acquired from the information distribution database 102, and the calculated number of browsing article items is calculated. By multiplying the feed importance of the recommended candidate feed to which the article item belongs, an article item of a predetermined value or more is finally extracted as an article item (feed) recommended to the user.

  Then, the recommended feed control unit 101b performs a feed aggregation process for aggregating the extracted article items in the same manner as the aggregated feed control unit 101a, and the aggregated recommended feed (aggregated recommended feed) is sent to the user terminal 200. be delivered.

  As described above, the recommended feed generation process according to the present embodiment distributes a highly useful feed that is highly likely to be of interest to the user and the user's user network to the user. In particular, in the recommended feed generation process of the present embodiment, the importance of feeds for users of a plurality of recommended candidate feeds is calculated, and further, the recommended candidate feeds are calculated based on the number of views of each article item included in the recommended candidate feeds. Each article item is extracted and generated as a recommendation feed. For this reason, even if the recommendation candidate feed has a low feed importance, if the number of article items included in the recommendation candidate feed is large, the importance of the article item to the user will be high, and the recommendation feed to the user Will be delivered as. In other words, the feed importance for the user of the recommendation candidate feed viewed by any other user other than the user network is low because the number of browsing in the user network is small, but is included in the recommendation candidate If there are a large number of browsing of each article item, it will be distributed to the user as a recommended feed. As described above, in this embodiment, the recommendation feed is generated based on the importance of each recommendation candidate feed for the user and the number of views of each article item included in the recommendation candidate feed, which is beneficial to the user. A highly recommended recommendation feed can be generated with high accuracy and based on a wide range of information. When the user browses the recommended feed, the browsing request process described above feeds back the browsing history, which affects the importance of the feed to other users in the user network. It is possible to obtain the same effect as “transfer the received information to another person”.

  10, FIG. 11 and FIG. 12 are examples of a WEB page provided to the user terminal 200 by the information distribution system of the present embodiment.

  The information distribution server 100 of the present embodiment posts various types of information generated by the above-described browsing request processing and recommended feed generation processing to each user who performed feed distribution registration on a WEB page as shown in FIGS. 10 and 11. To provide. This WEB page is recommended by the profile information of the user, the feed that the user has recently subscribed to, the most popular tag information for all users, the most popular feed that has been recently subscribed, and the recommended feed generation process. Weighted feed users (feed importance) and recommended feed articles are posted, and the user can register the newly recommended feed distribution from this WEB page or refer to the browsing history of the user Is possible.

  FIG. 12 is a WEB page on which the user network information of the user is posted. Information and similarity of each user in the user network automatically generated by the browsing request processing described above, and the user arbitrarily registers Information on each user in the intentional user network, tag information noted by the user himself (tag information based on the browsing history of the user), and tag information noted in the automatic or intentional user network It is possible. Further, as shown in FIG. 12, by marking each user of the automatically constructed user network, the marked user can be registered in the intentional user network.

  As described above, since the information distribution system of the present embodiment aggregates a plurality of feeds registered by the user into one and distributes them to the user terminal, even if the user does not individually acquire a feed from each feed, Only by registering the URL for feed distribution provided by the information distribution server in the feed reader, multiple feeds can be acquired at once, and each feed has undergone redirection processing. Even if there is no mechanism or user operation, it is possible to acquire browsing history (information selection, information browsing request, etc.) on the Internet that a user routinely performs.

  In particular, the information distribution system of this embodiment can construct a user network with similar preferences without being aware of the user by acquiring the browsing history for the feed. In other words, since a plurality of users associated with the feeds viewed by each user are built as a user network, the user's attributes are not limited to the network attributes, and the users are registered to feed users. Without overloading. It is possible to distribute and recommend highly useful information to each user.

Furthermore, the information distribution system according to the present embodiment includes an automatic user network constructed by users having similar preferences from a browsing history for a feed, and an intentional user network constructed by a user intentionally registering. One user network is provided.
For this reason, since the feed in the network intentionally configured by the user by the intentional user network and the feed in the automatic user network having a high preference similarity between the users are provided to the user, the user It will provide information selection, order, classification, etc. suitable for the taste.

  Furthermore, the information distribution system of this embodiment provides a recommended feed using tag information of each article item in the feed or feed. The tag information is keyword-like information representing a feed category or the like, and a plurality of feeds are associated with one tag information as described above. That is, the tag information which is interested in the user network is extracted, and among the plurality of feeds related to the tag information, a feed that the user in the user network has not yet browsed (a feed without a browsing history) It can be provided as a recommendation feed.

  This is because when a group of information that is interesting to a user is considered concentrically around a certain user, the central part of the user is feed distribution registration information registered by the user, and an intentional user network and automatic This is an information group in which each user in the target user network is interested, and further, on the outside, an information group in which the user and the user of the user network are not interested is spread. In such an information group, the information distribution system of this embodiment uses the tag information to associate the user with the feed, so that the user in the user network has never viewed the feed, that is, the user normally It is possible to provide a useful feed for the user from the information group that is not contacted. For this reason, it is possible to distribute a wide range of useful information to the user.

It is a schematic block diagram of the information delivery system of this invention. It is a block diagram of the configuration of the information distribution server of the present invention. It is a flowchart for demonstrating the process transition of the information delivery system of this invention. It is a flowchart for demonstrating the process transition of the information delivery system of this invention following FIG. 3A. It is a flowchart for demonstrating the process transition of the information delivery system of this invention following FIG. 3B. It is a flowchart for demonstrating the browsing request process in the information delivery system of this invention. It is a flowchart for demonstrating the recommendation feed production | generation process in the information delivery system of this invention. It is a flowchart for demonstrating the aggregation feed production | generation process in the information delivery system of this invention. It is a WEB page which the information delivery system of this invention provides, and is a figure which shows a feed delivery registration screen. It is a WEB page which the information delivery system of this invention provides, and is a figure which shows a user registration screen. It is a figure which shows the feed reader of the user terminal in the information delivery system of this invention. It is a figure for demonstrating the calculation method of the feed importance of the information delivery system of this invention. It is a WEB page provided by the information distribution system of the present invention. It is a WEB page provided by the information distribution system of the present invention. It is a WEB page provided by the information distribution system of the present invention.

Explanation of symbols

100 Information Distribution Server 101 Feed Distribution Unit 101a Aggregated Feed Control Unit 101b Recommended Feed Control Unit 102 Information Distribution Database 103 Feed Collection Unit 104 View Control Unit 105 Feed Importance Calculation Unit 106 User Network Construction Unit 107 Registration Control Unit 108 CPU
200 User terminal 300 WEB server (WEB site)

Claims (12)

An information distribution system including an information distribution server that collects a feed provided from a WEB site on the Internet and distributes the feed to a user terminal of a user,
The information distribution server is
Feed acquisition means for acquiring a plurality of feeds based on a plurality of feed distribution registration information registered via the user terminal from a WEB site on the Internet providing each feed;
Each feed distributed to the user terminal is subjected to redirect processing for connecting to each WEB site or the like that provides a feed via the information distribution server, and a plurality of the redirect processing is performed. An aggregate feed control means for generating an aggregate feed that aggregates the feeds;
When a browsing request for each feed of the aggregate feed is received from the user terminal, browsing control means for storing the browsing request for the feed as a browsing history of the user in a browsing history database;
User network construction means for generating user network information associating a plurality of users with similar browsing histories for each feed;
An information distribution system comprising: recommended feed control means for generating a recommended feed for recommending a feed to the user based on the browsing history of each user associated by the user network information.   The network construction means calculates a preference level of each user for each feed using the feed registration information and the browsing history, and a plurality of users whose browsing history for each feed is similar based on the calculated preference level The information distribution system according to claim 1, wherein the user network information associated with a user name is generated.   The information distribution system according to claim 1, wherein the browsing history is a browsing request for each feed in the aggregate feed or a browsing request for tag information associated with each feed.   The said recommendation feed control means produces | generates the said recommendation feed from the several feed browsed by each user linked | related by the said user network information, The one of Claim 1 characterized by the above-mentioned. Information distribution system.   The recommended feed control means extracts tag information included in each feed from a plurality of feeds browsed by each user associated by the user network information, and generates the recommended feed based on the extracted tag information. The information distribution system according to any one of claims 1 to 3.   The recommended feed control means calculates the importance for each feed of each user for each of a plurality of feeds browsed by each user associated by the user network information using the history information and the user network information, The information distribution system according to claim 1, wherein the recommended feed is generated based on importance of the feed. The information distribution server further includes user registration control means for registering other users who distribute information useful for the user or other users with similar preferences to the user from the user terminal,
The information distribution system according to any one of claims 1 to 6, wherein the user network construction unit generates intentional user network information in which other registered users are associated with each other.   The recommended feed control means performs redirection processing for connecting to each WEB site that provides a feed via the information distribution server to each of the generated recommended feeds, and the redirection processing is performed. The information distribution system according to claim 1, wherein an aggregate recommendation feed obtained by aggregating the plurality of feeds is generated. An information distribution system including an information distribution server that collects a feed provided from a WEB site on the Internet and distributes the feed to a user terminal of a user,
The information distribution server is
Feed acquisition means for acquiring a plurality of feeds based on a plurality of feed distribution registration information registered via the user terminal from a WEB site on the Internet providing each feed;
Each feed distributed to the user terminal is subjected to redirect processing for connecting to each WEB site or the like that provides a feed via the information distribution server, and a plurality of the redirect processing is performed. An aggregate feed control means for generating an aggregate feed that aggregates the feeds;
When a browsing request for each feed of the aggregate feed is received from the user terminal, browsing control means for storing the browsing request for the feed as a browsing history of the user in a browsing history database;
Recommended feed control means for generating a recommended feed based on the browsing history of another user who distributes information useful for the user registered from the user terminal or another user who has a preference similar to the user. An information distribution system characterized by this. An information distribution device that collects a feed provided from a WEB site on the Internet and distributes the feed to a user terminal of a user,
Feed acquisition means for acquiring a plurality of feeds based on a plurality of feed distribution registration information registered via the user terminal from a WEB site on the Internet providing each feed;
Each feed distributed to the user terminal is subjected to a redirect processing for connecting to each WEB site or the like that provides a feed via the information distribution device, and a plurality of the redirect processing is performed. An aggregate feed control means for generating an aggregate feed that aggregates the feeds;
When a browsing request for each feed of the aggregate feed is received from the user terminal, browsing control means for storing the browsing request for the feed as a browsing history of the user in a browsing history database;
User network construction means for generating user network information associating a plurality of users with similar browsing histories for each feed;
An information distribution apparatus comprising: a recommended feed control unit configured to generate a recommended feed for recommending a feed to the user based on the browsing history of each user associated by the user network information. In order to collect a feed provided from a web site on the Internet using a computer and distribute the feed to a user terminal of a user,
A first function for acquiring a plurality of feeds based on a plurality of feed distribution registration information registered via the user terminal from a WEB site on the Internet providing each feed;
Each feed distributed to the user terminal is subjected to redirect processing for connecting to each WEB site that provides the feed via the computer. And a second function for generating an aggregate feed obtained by aggregating the plurality of feeds subjected to the redirect processing,
A third function that, when receiving a browsing request for each feed of the aggregate feed from the user terminal, stores the browsing request for the feed as a browsing history of the user in a browsing history database;
A fourth function for generating user network information in which a plurality of users having similar browsing histories for each feed are associated;
And a fifth function for generating a recommended feed for recommending a feed to the user based on a browsing history of each user associated by the user network information. An information distribution method comprising an information distribution server for collecting a feed provided from a WEB site on the Internet and distributing the feed to a user terminal of a user, wherein the information distribution server includes:
A first step of acquiring a plurality of feeds based on a plurality of feed distribution registration information registered via the user terminal from a WEB site on the Internet providing each feed;
Each feed distributed to the user terminal is subjected to redirect processing for connecting to each WEB site or the like that provides a feed via the information distribution server, and a plurality of the redirect processing is performed. A second step of generating an aggregate feed that aggregates the feeds;
A third step of storing a browsing request for the feed in the browsing history database as a browsing history of the user when receiving a browsing request for each feed of the aggregate feed from the user terminal;
A fourth step of generating user network information associating a plurality of users with similar browsing histories for each feed;
And a fifth step of generating a recommended feed for recommending a feed to the user based on a browsing history of each user associated by the user network information.

Images