JP2008102846A - Content recommendation server, content recommendation program, and content recommendation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To recommend a content meeting user's taste without comparing past actions between individual users. <P>SOLUTION: A community similarity judgement part 124 compares a book mark list 113 of a first community and that of a second community with each other and obtains the degree of similarity by dividing the number of pieces of content specification information listed in respective book mark lists 113 by the number of pieces of content specification information listed in the book mark list 113 of the first community, and judges two lists to be similar when the degree of similarity exceeds a threshold. When two lists are determined to be similar, a recommended article distribution part 125 recommends content specification information which is listed in the book mark list 113 of the first community but is not listed in the book mark list 113 of the second community, to users belonging to the second community. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a content recommendation server, a content recommendation program, and a content recommendation method for recommending content that suits the user's preference (referred to as “user” as appropriate) from among a large number of contents.

There is known a technique for collaborative filtering that recommends a product that user B is likely to like based on user A's preference (estimates user B's preference using user A's data) (for example, patent literature). 1).
JP 09-265478 A

  Conventional collaborative filtering records each user's past behavior, and if the user's past behavior is similar to the past behavior of other users, the preference information of other users whose past behavior is similar Originally, the user's preference was guessed. For example, when there are 10,000 users, in order to estimate the preference of the user A, a user whose past behavior is similar to the user A is searched from among 9999 users other than the user A. For this reason, when the number of users increases, it is necessary to compare with past actions of many users before searching for a user with similar past actions.

  Therefore, an object of the present invention is to provide a content recommendation server, a content recommendation program, and a content recommendation method that can recommend content that suits user preferences without comparing past actions among individual users.

  The first feature of the present invention is a content recommendation server that recommends content that seems to meet the user's preference from among a large number of contents, each storing a user belonging to each community, and each community Storage means for storing a list of content specifying information registered due to a predetermined action of a user belonging to the community, a first community list and a second community list stored in the storage means And "the number of content specifying information listed in both the first community list and the second community list" is "the number of content specifying information listed in the first community list". A value obtained by dividing the total number by the similarity is a similarity, and when the similarity exceeds a threshold, it is determined that the two lists are similar. Content determination information that is listed in the first community list but not listed in the second community list when it is determined that the two lists are similar by the degree determination unit and the similarity determination unit. Content recommendation means for recommending to users belonging to the second community.

  “A predetermined action of a user belonging to the community” means, for example, “a user belonging to any community attaches a bookmark, writes a comment, or browses”. “Content” means articles such as bulletin board threads, blogs, news, and the like. “Content specifying information” means, for example, a URL that can specify an article or the like.

  “The similarity determination means determines that“ the number of content identification information listed in both the first community list and the second community list ”is“ the content identification listed in the first community list ”. The value divided by the “total number of information” is regarded as the similarity, and when the similarity exceeds a threshold, it is determined that the two lists are similar ”. For example, the list of the first community and the list of the second community When the number of content specifying information listed in both lists is “5”, the total number of content specifying information listed in the first community list is “10”, and the threshold is “0.49”, two The lists are determined to be similar.

“Recommend” means, for example, “send screen configuration information for displaying a list of content titles”. "Screen configuration information" is Opera (registered trademark) or Internet
It is described in HTML (Hyper Text Markup Language), CHTML (Compact HTML), etc. that can be browsed by a web browser such as Explorer (registered trademark).

  The second feature of the present invention further stores information that can specify the order in which the content specifying information is registered in the list of communities to which the content specifying information corresponds, and the similarity determination means includes m registered in the first community list. The similarity is determined based on the content identification information from the nth to the nth and the content identification information from the mth to the nth registered in the second community list.

“Information that can specify the order in which the content specifying information is registered in the list of the corresponding community” includes the time when the content specifying information is registered in the list. In addition, “content specifying information B is registered after the content specifying information A. The content specifying information C is registered after the content specifying information B” in “information that can specify the order in which the content specifying information is registered in the list”. Also included is information that can relatively specify the order in which each piece of content specifying information is registered.
In the second feature, a part of the content specifying information registered in the first community list is compared with a part of the content specifying information registered in the second community list. Calculate the similarity of the list. For example, the content specifying information registered immediately before the similarity determination is No. 1, the content specifying information registered before No. 1 is No. 2, and the content specifying information registered before No. 2 is No. 3. Content specified information registered in the first community list from No. 1 to No. 100 and content specified information registered immediately before similarity determination as No. 1 are registered before No. 1. The specific information is No. 2, the content specific information registered before No. 2 is No. 3, and the content specific information registered in the second community list No. 1 to No. 100 is compared with 2 Calculate the similarity of two lists.

  According to a third feature of the present invention, the storage unit further stores the registration time in association with the content specifying information, and the similarity determination unit stores the first time from the first time in the first community list. The similarity is calculated on the basis of the content specifying information registered in the period up to and the content specifying information registered in the second community list from the first time to the second time.

  The third feature is that a part of the content specifying information registered in the first community list is compared with a part of the content specifying information registered in the second community list. This is common to the second feature in that the similarity of the list is calculated. However, in the second feature, the similarity is calculated based on the content specifying information selected based on the order registered in the list. However, in the third feature, the content selected based on the time registered in the list. The similarity is calculated based on the specific information. For example, the content specifying information registered in the first list within 72 hours retroactively from the time when the similarity determination is executed, and registered in the second list within 72 hours similarly from the time when the similarity determination is executed. The similarity between the two lists is calculated by comparing with the specified content specifying information.

  According to a fourth feature of the present invention, the storage unit further stores a comment attached to the content specifying information, and the similarity determination unit is a content whose number of comments associated with the content specifying information exceeds a threshold value. The similarity is calculated based on the content specifying information.

  The fourth feature is that a part of the content specifying information registered in the first community list is compared with a part of the content specifying information registered in the second community list. This is common to the second feature in that the similarity of the list is calculated. However, in the second feature, the similarity is calculated based on the content specifying information selected based on the order registered in the list. In the fourth feature, the similarity is calculated based on the content specifying information selected based on the number of comments. Calculate the degree. For example, a list made up of content specifying information whose number of comments is registered in the first community list and added by a user belonging to the first community is five or more, and registered in the second list and second The similarity between the two lists is calculated by comparing with a list made up of content specifying information having five or more comments added by users belonging to the community.

  The fifth feature of the present invention is that the storage unit further stores the registration time in association with the content specifying information, and the content recommendation unit has an earlier registered time when recommending a plurality of content specifying information. The content specifying information is arranged in order or late.

  The sixth feature of the present invention is that the storage means also stores the registration time of a comment attached to the content identification information in association with the content identification information, and the content recommendation means recommends a plurality of content identification information. In this case, the content specifying information is arranged in order from the earliest or later time when the latest comment added to each content specifying information is registered.

  According to a seventh feature of the present invention, the storage means further stores a comment attached to the content specifying information, and the content recommendation means associates with the content specifying information when recommending a plurality of content specifying information. The content specifying information is arranged in order of increasing or decreasing number of comments.

  According to the first feature of the present invention, a list of content specifying information of two communities is compared, and when the similarity of the content specifying information registered in the list is high, it is registered in the list of one community. Content specifying information that is not registered in the list of the other community is presented to the user belonging to the other community. This makes it possible to determine the similarity between two communities based on the past behavior of one community and the past behavior of another community, and to infer the preference of the other community based on the preference of one community. It becomes possible.

  According to the second feature of the present invention, it is possible to determine the similarity between two lists based on a predetermined number of pieces of content specifying information extracted from each list.

  According to the third feature of the present invention, it is possible to determine the similarity between two lists based on content specifying information registered in each list within a predetermined period.

  According to the fourth aspect of the present invention, it is possible to determine the similarity between two lists based on content specifying information that is not only registered in the list but also has a predetermined number of comments in the community. It becomes.

  According to the fifth to seventh features of the present invention, it is easy for a user who receives recommended content specifying information to see recommended content specifying information.

  The best mode for carrying out the present invention will be described below with reference to the drawings. The following description is merely an example, and the technical scope of the present invention is not limited to the following description.

  In the content recommendation system according to the present embodiment, the content recommendation server recommends content that seems to meet the user's preference from among many contents.

  That is, although details will be described later, the content recommendation server stores information about the community. Specifically, a user belonging to the community is stored for each community. In addition, for each community, a list (bookmark list) including content specifying information registered by a user belonging to the community is stored.

  And (1) a list (first list) consisting of content specifying information registered by a user belonging to one community (first community) and a user belonging to another community (second community) A list (second list) made up of registered content specifying information is compared, and (2) “the number of content specifying information listed in both the first list and the second list” is set to “first”. The value divided by the “total number of content specifying information listed in one list” is used as the similarity, and (3) when the similarity exceeds the threshold, it is determined that the two lists are similar.

  Furthermore, when it is determined that the two lists are similar as a result of the similarity determination described above, content specifying information that is listed in the first list but not listed in the second list belongs to the second community. Recommend it to users.

  FIG. 1 shows an example of the overall configuration of a content recommendation system as an embodiment of the present invention. Hereinafter, the content is appropriately referred to as “article”.

  The browsing terminal 3 is a mobile phone owned by a user, for example, and performs wireless communication with the packet network 5 through a base station (not shown), and a plurality of article distribution servers 6 via the gateway server 4 and the Internet 2. Can be requested to browse articles stored in each article distribution server 6. The storage unit of the browsing terminal 3 stores an article browsing program for realizing article browsing processing. The article browsing program includes a Web browser. Articles are written in CHTML. The browsing terminal 3 includes a liquid crystal screen 31 that can display received articles and the like, and a key group 32 that is operated to select and click a button displayed on the liquid crystal screen 31.

  Similarly, the browsing terminal 3 can access the content recommendation server 1 and attach a bookmark to an article selected from the articles stored in the article distribution server 6. The bookmark is stored in the content recommendation server 1.

  FIG. 2 shows an example of the configuration of the content recommendation server 1 in the embodiment of the present invention. As shown in FIG. 2, the content recommendation server 1 records the storage unit 11, the processing unit 12, the communication unit 13, and the date and time when an article was posted or updated (hereinafter collectively referred to as “description”). The timepiece part 14 etc. used for the purpose etc. are provided.

  The storage unit 11 stores user information 111, community information 112, and a bookmark list 113. The storage unit 11 is, for example, an HDD (Hard Disk Drive), a RAM (Random Access Memory), or the like.

  The processing unit 12 includes a user registration unit 121, a community formation unit 122, a bookmark list formation unit 123, a community similarity determination unit 124, and a recommended article distribution unit 125.

  When a new user accesses the content recommendation server 1 by operating the key group 32 of the browsing terminal 3, the user registration unit 121 displays a screen for prompting input of user information on the liquid crystal screen 31 of the browsing terminal 3. The screen configuration information is transmitted to the browsing terminal 3 via the communication unit 13. Then, the user information 111 sent from the browsing terminal 3 is stored in the storage unit 11.

  The community forming unit 122 determines whether or not a new community can be formed, and determines whether or not a new member can participate in an existing community. And the community formation part 122 memorize | stores community ID, the user ID of the member of each community, etc. in the memory | storage part 11 as the community information 112. FIG.

  The bookmark list forming unit 123 forms, for each community, a bookmark list 113 that can be shared by a plurality of users belonging to the same community on a network such as the Internet. Specifically, a first bookmark list that can be written and browsed by members of the first community, a second bookmark list that can be written and browsed by members of the second community, and the like are formed.

  The community similarity determination unit 124 determines the similarity between communities. In the present embodiment, the similarity is determined based on the bookmark list of the community to be compared.

  The recommended article distribution unit 125 selects an article posted in the bookmark list of one community among the communities whose similarity is determined to satisfy the predetermined condition by the community similarity determination unit 124 as an article related to the other community (recommended Article).

  FIG. 3 shows an example of user information 111. As shown in FIG. 3, the user information 111 stores personal information such as a user name, e-mail address, age, birthday, and photograph, a community ID to which a user with the user ID belongs, and the like in association with the user ID. To do.

  FIG. 4 shows an example of the community information 112. As shown in FIG. 4, the community information 112 stores a community ID and a user ID of a member belonging to the community in association with each other.

  FIG. 5 shows an example of the bookmark list 113. As illustrated in FIG. 5, the bookmark list 113 stores a community ID and information for specifying an article with a bookmark (for example, Uniform Resource Locator: URL of the article) in association with each other. The bookmark list 113 further includes the user ID (for example, uid1) of the user who bookmarked the article (for example, article A), the date and time when the bookmark was registered (for example, 2006/08/01 22:31), the bookmark The title and content of the comment for the article with the comment, the user ID of the user who wrote the comment (for example, uid98), the date and time when the comment was written (for example, 2006/08/05 22:51), and Information (for example, URL) for specifying a related article may be included. In the above example, a title is assigned to a comment. However, a configuration that does not include a title may be used.

  FIG. 6 is a flowchart illustrating an example of a process flow in the content recommendation server 1. As shown in FIG. 6, all communities are executed as processing targets (loop A: steps S11 to S25).

  In loop A, first, another community to be compared is selected (step S13).

  In Step S15, the bookmark lists are compared between the comparison target communities (between the current processing target community and the other communities selected in Step S13), and the similarity is determined. If it is determined that the similarity satisfies the predetermined condition, that is, it is similar (YES in step S17), the process proceeds to step S19. If it is determined that the similarity does not satisfy the predetermined condition, that is, it is not similar (NO in step S17), the process proceeds to step S21.

  In step S19, articles that are not in the bookmark list of the community to be processed but are in the bookmark list of the comparison target community determined to be similar are extracted. The extracted article is stored in a RAM or the like.

  In step S <b> 21, it is determined whether or not an article has been extracted from all other communities as comparison targets for the processing target community. When articles are extracted with all other communities as comparison targets, the process proceeds to step S23, and when there is another community that is not compared, the process returns to step S13, and the next comparison target community is selected. In step S23, the extracted article is provided (distributed) as a related article to the community to be processed. If the end condition of loop A (the process has been executed for all communities) is satisfied, the process exits loop A and ends, and the process returns to step S11 if the end condition of loop A is not satisfied.

  The bookmark list similarity determination will be described. For example, each URL of an article registered in the bookmark list of the first community is compared with each URL of an article registered in the bookmark list of the second community. When there are many URLs registered in the bookmark list of the first community and also registered in the bookmark list of the second community, the similarity between the bookmark list of the first community and the bookmark list of the second community is similar. Determined to be high. Conversely, if there are few URLs registered in the bookmark list of the first community and also registered in the bookmark list of the second community, the bookmark list of the first community and the bookmark list of the second community are It is determined that the degree of similarity is low.

  For example, when more than half of the URLs registered in the bookmark list of the second community are also registered in the bookmark list of the first community, the bookmark list of the second community is the bookmark list of the first community. Is determined to be similar to

  In such a case, in step S19, the URL registered in the bookmark list of the second community but not registered in the bookmark list of the first community is used as the URL of the recommended content for the first community. As offered.

  When determining the similarity of the bookmark list, the date and time when the bookmark was attached may be considered. For example, the bookmark list of the first community registered within a predetermined number of days (for example, 7 days) from the date for determining similarity, and registered within the predetermined number of days (for example, 7 days) from the date for determining the similarity If the duplication rate of the URLs registered in the bookmark lists of the two communities is equal to or greater than a threshold (for example, 0.5) It is determined that the lists are similar.

  Needless to say, the time may be used as a reference instead of the date. For example, the bookmark list of the first community registered within a predetermined time (for example, 72 hours) from the time when the similarity determination is performed, and registration within the predetermined time (for example, 72 hours) from the time when the similarity determination is performed. If the duplication rate of URLs registered in the bookmark lists of the two communities is equal to or greater than a threshold (for example, 0.4) by comparing with the bookmark list of the second community that has been set, the bookmarks of the two communities It is determined that the lists are similar.

  Further, the order of registration in the bookmark list may be taken into consideration when determining the similarity of the bookmark list. For example, the content specifying information registered immediately before the similarity determination is number 1, the content specifying information registered before number 1 is number 2, and the content specifying information registered before number 3 is number 3. Content specifying information registered in the first list from No. 1 to No. 100 and content specifying information registered immediately before the similarity determination is set to No. 1, and content specifying information registered before No. 1 2 and the content identification information registered before 3 as the third, and the content identification information registered in the second list from the 1st to the 100th is compared with the similarity between the two lists Calculate the degree. And when the duplication rate of URL registered in the bookmark list of two communities is more than a threshold value (for example, 0.6), it determines with the bookmark list of two communities being similar.

  In addition, when recommending an article URL from one community to another community among communities determined to have similar bookmark lists, articles registered in one bookmark list but not registered in the other bookmark list A part of the URL may be recommended. For example, (1) recommend only articles whose number of comments is equal to or greater than a predetermined number, (2) recommend only articles whose date the article URL is registered in the bookmark list is within a predetermined number of days from the date when the similarity is determined. (3) We recommend only articles whose time when the article URL is registered in the bookmark list is within a predetermined time from the time when the similarity is determined.

  Based on FIG. 7, an overview of processing in the embodiment of the present invention will be described. As shown in FIG. 7 as an example, a community C1 is formed by three users M11, M12, and M13. In the community C1, communication is performed by attaching a bookmark to an article that can be browsed on the Internet and writing a comment on the article with the bookmark.

In FIG. 7, user M11 bookmarks article A,
Next, the user M12 writes a comment a2 on the article A,
Next, the user M13 writes a comment a3 on the article A,
Next, the user M11 writes a comment a4 on the article A,
Next, the user M11 further writes a comment a5 on the article A,
Next, the user M13 writes a comment a6 on the article A,
Next, the user M12 writes a comment a7 on the article A,
Next, the user M11, the user M12, and the user M13 communicate with each other by, for example, writing a comment a8 on the article A.

  Similarly, in the community C1, someone who is a member bookmarks the article B, the article C, the article G, and the article R, and then writes a comment. Although not shown in the figure, in another community C2, a member of the community C2 bookmarks Article A, Article B, Article C, Article D, Article E, Article F, and Article G, and then writes a comment. It is out.

  Information on bookmarks and comments written in the community is stored as a bookmark list 113 for each community. Further, the following processing is executed for each community at a predetermined timing (for example, a predetermined time / predetermined time interval).

  First, the bookmark list BL1 of the community C1 and the bookmark list BL2 of the community C2 are compared to determine the similarity (step S15 in FIG. 6). Specifically, the bookmark list BL1 is composed of URLs of five articles, Article A, Article B, Article C, Article G, and Article R, and the bookmark list BL2 is Article A, Article B, Article C, Article D, Article. It consists of the URLs of seven articles, E, Article F, and Article G. The URLs of the four articles, Article A, Article B, Article C, and Article G, are common to the two bookmark lists. That is, the URLs of four articles, more than half of the seven article URLs registered in the bookmark list BL2, are also registered in the bookmark list BL1. The threshold used for similarity determination is 0.5.

  Therefore, in step S17 of FIG. 6, it is determined that the bookmark list BL1 and the bookmark list BL2 are similar, and the process proceeds to step S19.

  In step S19, the article D, the article E, and the article F not in the bookmark list BL1 but in the bookmark list BL2 are provided as related articles (recommended articles) from the community C2 to the community C1.

  FIG. 8 shows an example of a screen on which related articles (recommended articles) are displayed. The screen shown in FIG. 8 is a screen that can be browsed by members of the community C1. Titles 811, 821 and 831 are titles of articles registered in the bookmark list of the community C1 by any member of the community C1. The sentence 813 is the content of the article corresponding to the title 811, the sentence 823 is the title 821, and the sentence 833 is the content of the article corresponding to the title 831. The title list 815 is a sentence 813, the title list 825 is a sentence 823, and the title list 835 is a title of a comment written by any member of the community C1 to the sentence 833. The title lists 815, 825, and 835 are displayed together with the date and time when each comment was written. In the member list 841, the handle names of the members of the community C1 are displayed. The recommended article title list 851 displays the titles of recommended articles.

(Modification)
Of articles viewed in the community (articles browsed by members of the community), register an article with the number of browsing more than a predetermined number in the interested list, and determine the similarity between communities based on the similarity of the interested list It is also good. For example, as shown in FIG. 9A, for each community, data of articles browsed in the community is updated and stored as browsing history information 115. The browsing history information 115 stores the community ID, the URL of the viewed article, the user ID of the browsing user, and the browsing date and time. If there is an article whose number of browsing is equal to or greater than a predetermined number, the article is registered in the interest list 117 shown in FIG. In this interest list 117, community IDs and URLs of articles whose number of browsing is equal to or greater than a predetermined number are stored.

  Note that the browsing history may be stored and stored for each user instead of storing the browsing history for each community. For example, as shown in FIG. 10A, for each user, the data of articles browsed by the user is updated and stored as browsing history information 116. The browsing history information 116 stores a user ID, a URL of a browsed article, a browsing time, and an interest flag indicating that the browsing count has reached a predetermined number or more. If there is an article whose number of browsing is equal to or greater than a predetermined number, the article is registered in the interest list 117 shown in FIG.

  When an article registered in the interest list 117 is updated, the corresponding community member is notified that the article has been updated, and another community in which the interest list 117 is determined to be similar. The updated article may be provided as a related article (recommended article).

  Techniques for checking whether registered articles have been updated include RSS readers, WWWDs, and keyword checkers. The RSS reader is software that collects RSS files that describe the update status and outline of the website in a lump and checks the update status of each website. WWWD (World Wide Web Dispatcher) is software that can visit a specified website and obtain the update date and time. The keyword checker is software that can notify articles including keywords that are set in advance.

The block diagram which shows an example of the whole structure of the content recommendation system as embodiment of this invention. The block diagram which shows an example of a structure of the content recommendation server 1 in embodiment of this invention. The figure which shows an example of the user information 111. The figure which shows an example of the community information 112. FIG. The figure which shows an example of the bookmark list | wrist 113. FIG. 5 is a flowchart showing an example of a process flow in the content recommendation server 1. The figure which shows the whole image of the process in embodiment of this invention. The figure which shows an example of the screen on which the recommendation article in embodiment of this invention is displayed. The figure which shows an example of the browsing history information and interest list in the modification of this invention. The figure which shows the other example of the browsing history information in the modification of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Content recommendation server 2 ... Internet 3 ... Browse terminal 4 ... Gateway server 5 ... Packet network 6 ... Article distribution server 11 ... Storage part 12 ... Processing part 13 ... Communication part 14 ... Clock part 31 ... Liquid crystal screen 32 ... Key group

Claims (9)

It is a content recommendation server that recommends content that seems to meet the user's preference among many content,
Storage means for storing each user belonging to each community, and storing a list of content specifying information registered for each community due to a predetermined action of the user belonging to the community;
The first community list and the second community list stored in the storage means are compared, and “contents listed in both the first community list and the second community list” are identified. Similarity determined by dividing the number of information by the “total number of content specifying information listed in the first community list” as similarity, and determining that the two lists are similar when the similarity exceeds a threshold Degree determination means;
When it is determined by the similarity determination means that the two lists are similar, the content specifying information that is listed in the first community list but not listed in the second community list is stored in the second community. Content recommendation means to recommend to users belonging to
A content recommendation server comprising: The storage means further stores information capable of specifying the order in which the content specifying information is registered in the list of the corresponding community,
The similarity determination means includes the mth to nth content specifying information registered in the first community list and the mth to nth content specifying information registered in the second community list. The content recommendation server according to claim 1, wherein similarity is determined based on (m and n are arbitrary natural numbers). The storage means further stores the registration time in association with the content specifying information,
The similarity determination means includes the content specifying information registered in the first community list from the first time to the second time, and the second community list from the first time to the second time. The content recommendation server according to claim 1, wherein the similarity is calculated based on content specifying information registered up to the time. The storage means further stores a comment attached to the content specifying information,
4. The similarity determination unit according to claim 1, wherein the similarity determination unit calculates the similarity based on content specifying information in which the number of comments associated with the content specifying information exceeds a threshold. Content recommendation server. The storage means further stores the registration time in association with the content specifying information,
5. The content recommendation unit arranges the content specification information in order from the earliest or later time of registration when recommending a plurality of pieces of content specification information. 6. Content recommendation server. The storage means further stores a registration time of a comment attached to the content specifying information in association with the content specifying information,
6. The content recommendation unit, when recommending a plurality of pieces of content identification information, arranges the latest comments attached to each piece of content identification information in order from the earliest to the latest registration time. The content recommendation server according to any one of the above. The storage means further stores a comment attached to the content specifying information,
The content recommendation unit arranges a plurality of pieces of content specifying information in order of increasing or decreasing number of comments associated with each piece of content specifying information. The content recommendation server described in the section. Each storage unit stores a user who belongs to each community, and includes a storage unit that stores a list of content identification information registered due to a predetermined action of a user belonging to the community for each community. A content recommendation server computer that recommends content that seems to meet user preferences from
The first community list and the second community list stored in the storage means are compared, and “contents listed in both the first community list and the second community list” are identified. Similarity determined by dividing the number of information by the “total number of content specifying information listed in the first community list” as similarity, and determining that the two lists are similar when the similarity exceeds a threshold Degree determination means,
When it is determined by the similarity determination means that the two lists are similar, the content specifying information that is listed in the first community list but not listed in the second community list is stored in the second community. Content recommendation means to recommend to users belonging to
Content recommendation program to make it function as Each storage unit stores a user who belongs to each community, and includes a storage unit that stores a list of content identification information registered due to a predetermined action of a user belonging to the community for each community. A content recommendation method in a content recommendation server that recommends content that seems to meet the user's preference from among,
The list of the first community and the list of the second community are compared, and “the number of content specifying information listed in both the first community list and the second community list” is set to “the first community list”. A value divided by the “total number of content specifying information listed in the community list” as a similarity, and determining that the two lists are similar when the similarity exceeds a threshold;
When it is determined that the two lists are similar, the content specifying information that is listed in the first community list but not listed in the second community list is sent to users belonging to the second community. Recommending steps,
A content recommendation method comprising:

Images