JP2005165454A - Content recommendation device, content recommendation method, and content recommendation program - Google Patents

Abstract

PROBLEM TO BE SOLVED To recommend an optimum content according to a user's situation at the time when a user accesses the content distribution service without burdening the user in the content distribution service.
A content recommendation system 100 includes a client terminal 1 used by a user and a server apparatus 2 connected to the client terminal 1 via a computer network. In such a system configuration, the server device 2 is a content distribution device that distributes broadcast and VOD type content distribution services represented by digital broadcasting to the client terminal 1, and the client terminal 1 Enjoy the content distribution service. In addition, the client terminal 1 uses the content metadata distributed from the server device 2 to present a recommendation of the content that the user will view to the user in consideration of the user's preference according to the user's situation. It is a content recommendation device.
[Selection] Figure 1

Description

  The present invention provides a content recommendation method for recommending a content to a user by using metadata describing the content of the content distributed and broadcast together with the content in a content distribution service represented by digital broadcasting, and content recommendation The present invention relates to a method and a content recommendation program.

Examples of content guide systems for content distribution services that distribute multimedia content such as video and music include the following. For example, in an “electronic program guide display device and display method” (see, for example, Patent Document 1), which is a content display system based on a genre, the content is classified by genre and a content table of only the genre selected by the user is displayed. To do. In addition, in a “question-based electronic program guide” (see, for example, Patent Document 2), which is a content presentation system using a user profile, the system asks the user a simple question in advance, and the preference etc. Create and save user profiles, and when the user views the content, first filter the content with the user profile, then the system asks the user a detailed question and presents the content selected by the system according to the answer . Furthermore, in “a program guide device, an inference device and an inference method” (for example, see Patent Document 3), which is a content presentation system using a user's content viewing history, the user's past viewed content names and future broadcasts The content names of the scheduled content are compared, and the common content is preferentially displayed as content that meets the user's preference.
JP 1998-191193 A Japanese Patent Laid-Open No. 1999-8810 JP 2001-275056 A

  However, the above-described prior art has the following problems.

  The “electronic program guide display device and display method” is a simple method, but there is a problem that as the number of contents increases, the content of the same genre increases, and the user has a burden to select from the list.

  The “question-based electronic program guide” allows users to control the presentation of large amounts of content by answering system questions when accessed. However, since updating the user profile is not taken into account, it may not be possible to respond to changes in the user's fundamental interests, or the same question may be repeated from the system, in which case there is a problem that the burden on the user is large .

  “Program guide device, reasoning device and reasoning method” can reduce the burden on the user and can cope with long-term changes in the user's interests. There is a problem that it cannot respond to changes in demand.

  The present invention has been made in view of the above circumstances, and in a content distribution service, content that recommends optimum content according to the user's situation at the time when the user accesses the content distribution service without imposing a burden on the user. An object is to provide a recommendation device, a content recommendation method, and a content recommendation program.

  To achieve the above object, the present invention according to claim 1 is capable of communicating with each other via a computer network and a content distribution apparatus that distributes a plurality of contents, and contents of the content acquired from the content distribution apparatus. A content recommendation device that recommends content that a user will view using metadata to be described, and stores the content viewing history of the user as first preference information for each viewing situation Separate user profile storage means, long-term user profile storage means for storing the user's content viewing history as second preference information for each user regardless of viewing status, and the user recommending the content recommendation Viewing status acquisition means for acquiring the viewing status of the user when the device is accessed; The metadata acquisition means for acquiring the metadata of the content, and for each of the acquired metadata of the plurality of contents, the viewing situation corresponding to or similar to the viewing situation acquired by the viewing situation acquisition means First preference information is acquired from the situation-specific user profile storage means, the acquired first preference information and the second preference information and the relevance between the contents obtained from the metadata, and Based on the respective contributions of the first and second preference information, a recommendation degree calculating means for calculating a score relating to content recommendation, and rearranging the scores for each content, to extract the content to be recommended, It has a gist of having recommendation result display means for displaying to the user.

  According to a second aspect of the present invention, in the first aspect of the present invention, the content-selected information selected by the user from the content to be recommended displayed to the user is stored in the user profile storage unit according to the situation. This is reflected in the first preference information and the second preference information in the long-term user profile storage means.

  The third aspect of the present invention is the common aspect of the first or second aspect of the present invention, wherein third preference information obtained by calculating the content viewing history of all the users for each viewing situation is acquired from the content distribution apparatus. It has a profile acquisition unit, and the recommendation level calculation unit further takes into account the third preference information and the degree of contribution of the third preference information.

  According to a fourth aspect of the present invention, in the invention according to any one of the first to third aspects, the content is selected when the user selects content from the content to be recommended displayed to the user. Viewing status storage means for storing the viewing status that has influenced the selection, wherein the viewing status acquisition means acquires one or a plurality of viewing statuses, and displays the viewing status that matches or is similar to the acquired viewing status. The summary is obtained from the viewing situation storage means, and the recommendation degree calculation means obtains the first preference information according to the viewing situation obtained from the viewing situation storage means.

  The present invention according to claim 5 is the invention according to any one of claims 1 to 4, wherein the preference information is used for each genre classified into a content category and a word indicating content content. The higher the user's preference, the higher the score, and the higher the score, the metadata has a predetermined score value for each genre when the content corresponds to the genre. .

  The present invention according to claim 6 is the invention according to any one of claims 1 to 5, wherein the viewing status is at least one of time information and position information, schedule information, and clothes information of the user. It has a gist.

  According to the seventh aspect of the present invention, there is provided metadata that describes a content content acquired from the content distribution device by a content recommendation device that can communicate with the content distribution device that distributes a plurality of contents via a computer network. A content recommendation method for recommending content that a user will watch using, and storing the content viewing history of the user in the storage unit as first preference information for each viewing situation A user profile storage step, a long-term user profile storage step of storing the user's content viewing history in the storage unit as second preference information for each user regardless of the viewing status, and the user A viewing status acquisition step for acquiring the viewing status of the user when accessing the content recommendation device. A metadata acquisition step for acquiring the metadata of the plurality of contents, and the viewing situation acquired in the viewing status acquisition step for each of the acquired metadata of the plurality of contents. The first preference information corresponding to the viewing situation to be obtained is acquired from the storage unit, and the obtained first preference information and the second preference information are related to the content content obtained from the metadata. And a recommendation degree calculating step for calculating a score relating to content recommendation based on the respective contribution degrees of the first and second preference information, and extracting the content to be recommended by rearranging the scores for each content And a recommendation result display step for displaying to the user.

  According to an eighth aspect of the present invention, in the seventh aspect of the present invention, the content selection information selected by the user from the content to be recommended displayed to the user is stored in the first storage unit. The gist is that it is reflected in the preference information and the second preference information.

  According to a ninth aspect of the present invention, in the seventh or eighth aspect of the present invention, the third preference information obtained by calculating the content viewing history of all the users for each viewing situation is obtained from the content distribution apparatus. The present invention includes a profile acquisition step, and the recommendation level calculation step further includes the third preference information and the contribution degree of the third preference information.

  According to a tenth aspect of the present invention, in the invention according to any one of the seventh to ninth aspects, the content is selected when the user selects content from the content to be recommended displayed to the user. A viewing status storage step for storing the viewing status that has influenced the selection in the viewing status storage unit, wherein the viewing status acquisition step acquires one or a plurality of viewing statuses and matches the acquired viewing status or The gist is that a similar viewing situation is obtained from the viewing situation storage unit, and the recommendation degree calculating step obtains the first preference information according to the viewing situation obtained from the viewing situation storage unit.

  The present invention according to claim 11 is the invention according to any one of claims 7 to 10, wherein the preference information is used for each genre classified into words indicating contents category and contents. The higher the user's preference, the higher the score, and the higher the score, the metadata has a predetermined score value for each genre when the content corresponds to the genre. .

  According to a twelfth aspect of the present invention, in the invention according to any one of the seventh to eleventh aspects, the viewing status is at least one of time information and position information, schedule information, and clothes information of the user. It has a gist.

  According to a thirteenth aspect of the present invention, a computer that can communicate with a content distribution apparatus that distributes a plurality of contents via a computer network uses metadata that describes the contents of the content acquired from the content distribution apparatus. And a content recommendation program for recommending content that the user will view, wherein the content viewing history of the user is stored in the storage unit as first preference information for each viewing situation. A profile storage step, a long-term user profile storage step in which the content viewing history of the user is stored in the storage unit as second preference information for each user regardless of the viewing status; and Viewing status acquisition that acquires the viewing status of the user when the recommended device is accessed And the metadata acquisition step for acquiring the metadata of the plurality of contents, and for each of the acquired metadata of the plurality of contents, match or similar to the viewing situation acquired in the viewing situation acquisition step The first preference information corresponding to the viewing situation is acquired from the storage unit, and the relationship between the acquired first preference information and the second preference information and the content content obtained from the metadata, And a recommendation degree calculating step for calculating a score relating to content recommendation based on the respective contribution degrees of the first and second preference information, and extracting the content to be recommended by rearranging the scores for each content. The gist is to cause the computer to execute a recommendation result display step to be displayed to the user.

  According to a fourteenth aspect of the present invention, in the invention according to the thirteenth aspect, content selection information selected by the user from content to be recommended displayed to the user is stored in the first storage unit. The gist is that it is reflected in the preference information and the second preference information.

According to a fifteenth aspect of the present invention, in the invention according to the thirteenth or fourteenth aspect, the third preference information obtained by calculating the content viewing history of all the users for each viewing situation is obtained from the content distribution apparatus. Causing the computer to execute a profile acquisition step;
The gist of the recommendation level calculating step is to further consider the third preference information and the contribution degree of the third preference information.

  The present invention according to claim 16 is the invention according to any one of claims 13 to 15, wherein the content is selected when the user selects content from the content to be recommended displayed to the user. The computer is caused to execute a viewing status storage step of storing the viewing status affecting the selection in the viewing status storage unit, and the viewing status acquisition step acquires one or a plurality of viewing statuses and the acquired viewing status Is obtained from the viewing situation storage unit, and the recommendation degree calculating step obtains the first preference information according to the viewing situation obtained from the viewing situation storage unit. .

  The invention according to claim 17 is the invention according to any one of claims 13 to 16, wherein the preference information is used for each genre classified into a content category and a word indicating the content. The higher the user's preference, the higher the score, and the higher the score, the metadata has a predetermined score value for each genre when the content corresponds to the genre. .

  The present invention according to claim 18 is the invention according to any one of claims 13 to 17, wherein the viewing status is at least one of time information and position information, schedule information, and clothes information of the user. It has a gist.

  According to the present invention, in a content distribution service, a content recommendation device, a content recommendation method, and a content recommendation method for recommending optimum content according to a user's situation at the time when the user accesses the content distribution service without burdening the user, and A content recommendation program can be provided.

  Thereby, the user can easily select and browse the optimum content for his current situation.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<First Embodiment>
FIG. 1 is a schematic configuration diagram of a content recommendation system 100 according to an embodiment of the present invention. A content recommendation system 100 shown in FIG. 1 includes a client terminal 1 used by a user and a server device 2 connected to the client terminal 1 via a computer network. In such a system configuration, the server device 2 is a content distribution device that distributes a broadcast and VOD (Video On Demand) type content distribution service typified by digital broadcasting to the client terminal 1. The terminal 1 can enjoy the content distribution service. In addition, the client terminal 1 uses the content metadata distributed from the server device 2 to present a recommendation of the content that the user will view to the user in consideration of the user's preference according to the user's situation. It is a content recommendation device.

  More specifically, the client terminal 1 includes an access detection unit 11, a content acquisition unit 12, a user status acquisition unit 13, a content priority acquisition unit 14, a recommendation result generation unit 15, a recommendation result display unit 16, a content display unit 17, and a content The selection action acquisition unit 18, the content selection action registration unit 19, the profile generation unit 110, and the profile database 120 are provided.

  The access detection unit 11 detects a user's access (for example, power-on to the client terminal 1 and the like), and transmits access information to the content acquisition unit 12 and the user status acquisition unit 13 when there is an access. .

  When the access information of the user is sent from the access detection unit 11, the content acquisition unit 12 sends metadata about the current and future content to the metadata distribution unit 21 of the server device 2 (for example, specifically, at least, Requesting a broadcast channel, broadcast time, program name, and program content), and sending the returned metadata to the content priority acquisition unit 14.

  Upon receiving the user access information from the access detection unit 11, the user status acquisition unit 13 acquires information indicating the user status (for example, time information in the case of acquisition from a clock or the like), and a content priority acquisition unit 14 is transmitted.

  Upon receiving the user status from the user status acquisition unit 13, the content priority acquisition unit 14 receives a user profile 121 by status that is a user profile that matches or is closest to the status from the profile database 120 and a long-term user profile unique to the user. A typical user profile 122 is acquired, the content recommendation level is calculated together with the content metadata obtained from the content acquisition unit 12, and is transmitted to the recommendation result generation unit 15.

  Here, the user profile is data representing the characteristics of the user's content viewing, and the situation-specific user profile 121 and the long-term user profile 122 are stored in the profile database 120. Configuration examples of the situation-specific user profile 121 and the long-term user profile 122 are shown in FIGS. The situation-specific user profile 121 illustrated in FIG. 2 includes content viewing score values (weights) for each of a plurality of situations for the same user, and more specifically, categories (sports, news, drama, etc.) and Score values are provided for each genre classified into words (for example, Matsui, Iraq, National Diet). Further, the long-term user profile 122 shown in FIG. 3 includes score values (weights) of content viewing for each user that do not depend on a specific situation (including all situations). , News, drama, etc.) and words (for example, Matsui, Iraq, Diet, etc.), score values are provided for each genre. The score value takes a value in the range of [0, 1]. The larger the value, the more the content is viewed.

A method for calculating the recommendation level of content will be specifically described. User i, content (metadata) j, situation k, number of genres m, score value of user profile 121 by situation is P (i, k), score value of long-term user profile 122 is L (i), and content If the score value of j is expressed as C (j),
P (i, k) = (P _ik1 , P _ik2 , ..., P _ikt , ..., P _ikm )
L (i) = (L _i1 , L _i2 , ..., L _it , ..., L _im )
C (j) = ( _Cj1 , _Cj2 ..., _Cjt , ..., _Cjm )
It becomes. Here, P _ikt and L _it are values in the range of [0, 1] representing weights (score values) relating to the genre t of the user profile 121 and long-term user profile 122, respectively, and C _jt is the content C ( j) is any value of {0, 1} indicating whether or not it is classified into the genre t.

  The following method is shown as an example of content recommendation score calculation.

(1) Method of expressing score by sum of situation-specific user profile 121 and long-term user profile 122 In this case, score value S (i, j, k) of score recommendation degree of content j of user i in situation k Is represented by the following equation.

However, α _i represents the degree of contribution of the user profile 121 according to the situation of the user i (an index indicating how much importance is placed on the situation-specific profile 121 or the long-term user profile 122 and is set in advance). It is a value in the range of [0, 1] that is sequentially updated by selecting the content. Further, the score value S (i, j, k) of the content recommendation level is calculated for all genres t where C _jt = 1 holds.

  When this method is used, the score value S (i, j, k) of the content recommendation level becomes higher as the content matches the user preference expressed by the situation-specific user profile 121 or the long-term user profile 122.

(2) A method of expressing the score by the product of the situation-specific user profile 121 and the long-term user profile 122.

  If this method is used, the score value S (i, j, k) of the content recommendation level has a high score for content that has a common user preference in both the situation-specific user profile 121 and the long-term user profile 122.

  Note that the method for obtaining the score value of the content recommendation level is not limited to the above-described score calculation, and other methods for obtaining the score are also possible. In addition, other information can be used for the data items included in the user profile, and the score may be obtained using the information.

  The recommendation result generation unit 15 arranges the contents in the order of score from the content recommendation score values obtained from the content priority acquisition unit 14 and sends them to the recommendation result display unit 16.

  The recommendation result display unit 16 presents a list of recommended content to the user based on the content recommendation score value sent from the recommendation result generation unit 15 (for example, content having a score value equal to or higher than a predetermined threshold, predetermined content) The content up to the top of the content) is sent to the content display unit 17 and the content selection action acquisition unit 18.

  The content display unit 17 uses the content list sent from the recommendation result display unit 16 to issue a content delivery request to the content delivery unit 23 and displays the content sent from the content delivery unit 23. ing.

  The content selection behavior acquisition unit 18 acquires the user's selection behavior for the content list presented to the user by the recommendation result display unit 16 (what content the user has selected), and the content selection behavior registration unit 19 It is supposed to send.

  The content selection behavior registration unit 19 registers the user's content selection behavior sent from the content selection behavior acquisition unit 18 in the content selection behavior database 22 of the server device 2.

  The profile generation unit 110 generates a situation-specific user profile 121 and a long-term user profile 122 from the content selection behavior database 22 and registers them in the profile database 120. That is, the user's content selection behavior is reflected in the score value (weight) of the profile database 120.

  The client terminal 1 and the server device 2 are electronic devices having a main storage device (memory) including a central processing unit (CPU) having at least a calculation function and a control function, and a RAM having a function of storing programs and data. It consists of a typical device. Each device may include an auxiliary storage device such as a hard disk in addition to the main storage device.

  Among these, the access detection unit 11, the content acquisition unit 12, the user status acquisition unit 13, the content priority acquisition unit 14, the recommendation result generation unit 15, the recommendation result display unit 16, the content display unit 17, and the content selection behavior of the client terminal 1. The acquisition unit 18, the content selection action registration unit 19, the profile generation unit 110, the metadata distribution unit 21 and the content distribution unit 23 of the server device 2 are other than those specifically showing the calculation control function by the CPU. Don't be. Further, the profile database 120 of the client terminal 1 and the content selection behavior database 22 of the server device 2 have functions of a main storage device.

  A program for executing various processes according to the present embodiment is stored in the main storage device or the hard disk described above. The program can be recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, or a DVD-ROM, or can be distributed via a communication network.

  Next, the operation of the content recommendation system 100 according to the present embodiment will be described using FIG. 4 and FIG. Here, FIG. 4 is a flowchart showing the content recommendation operation in the content recommendation system 100, and FIG. 5 is a flowchart showing the operation in which the content selection behavior of the user is reflected in the profile database 120.

  First, the content recommendation operation when the user accesses the content recommendation system 100 will be described. The access detection unit 11 of the client terminal 1 detects the user's access and transmits the user ID to the content acquisition unit 12 and the user status acquisition unit 13 (step S110).

  Next, the user status acquisition unit 13 acquires the user status (for example, time information) and transmits it to the content priority acquisition unit 14 (step S120). In addition, the content acquisition unit 12 requests content data (content metadata) from the metadata distribution unit 21, and transmits the returned content data to the content priority acquisition unit 14 (step S130).

  Next, the content priority acquisition unit 14 acquires the relevant user profile 121 and long-term user profile 122 from the profile database 120 based on the acquired user ID and the user status, and the score value of the content recommendation level Is transmitted to the recommendation result generation unit 15 (step S140). The recommendation result generation unit 15 rearranges the content in order of score based on the calculated score value of the recommendation level of each content, and transmits it to the recommendation result display unit 16 (step S150).

  Next, the recommendation result display unit 16 displays a list of recommended contents as a recommendation result, and transmits the content selected by the user to the content display unit 17 and the content selection action acquisition unit 18 (step S160).

  Next, the content selection behavior acquisition unit 18 transmits the acquired content selection behavior to the content selection behavior registration unit 19, and the content selection registration unit 19 registers the content selection behavior in the content selection behavior database 22 (step S170). ).

  Next, an operation in which a user's content selection behavior is reflected in the profile database 120 will be described. The content selection database 22 periodically transmits the user's content selection behavior to the profile generation unit 110 of the client terminal 1 according to the user ID from the registered data (step S210).

  The profile generation unit 110 generates a long-term user profile 122 that reflects all user content selection behaviors, regardless of the situation, and registers it in the profile database 120 (step S220).

  Further, the profile generation unit 110 classifies the user's content selection behavior according to the situation, generates a situation-specific user profile 121 reflecting the content selection behavior for each situation, and registers it in the profile database 120 (step S230).

  Therefore, according to the present embodiment, the user's situation is acquired, and the content recommendation is performed in consideration of the user profile based on the user's situation, so that the user's situation is met without burdening the user. The best content can be recommended.

  Specifically, for example, it can be understood that a certain user likes a drama or a baseball game from the past viewing history, but since a long time content is not viewed in the morning, a baseball game result in the news The news program is recommended and presented.

  In the first embodiment, the content selection behavior is registered in the content selection database 22 of the server device 2 and then reflected in the profile database 120 on the client terminal 1. Instead of transmitting to the apparatus 2, processing may be performed only on the client side and reflected in the profile database 120 as needed.

<Second Embodiment>
FIG. 6 is a schematic configuration diagram of a content recommendation system 200 according to the second embodiment of the present invention. The content recommendation system 200 shown in FIG. 6 has a configuration including a client terminal 3 used by a user and a server device 4 connected to the client terminal 3 via a computer network. In such a system configuration, the server device 4 is a content distribution device that distributes broadcast and VOD type content distribution services represented by digital broadcasting to the client terminal 3, and the client terminal 3 The content distribution service can be enjoyed. In addition, the client terminal 3 uses the content metadata distributed from the server device 4 to present to the user a recommendation of content that the user will view, taking into account the user's preference according to the user's situation It is a content recommendation device. Furthermore, in the second embodiment, content recommendation is performed in consideration of other users' preferences. In the present embodiment, only the configuration and functions different from those of the first embodiment will be described, and the other components and functions will be denoted by the same reference numerals and the description thereof will be omitted.

  More specifically, the client terminal 3 includes an access detection unit 11, a content acquisition unit 12, a user status acquisition unit 13, a content priority acquisition unit 31, a recommendation result generation unit 15, a recommendation result display unit 16, a content display unit 17, and a content The selection action acquisition unit 18, the content selection action registration unit 19, the profile generation unit 110, and the profile database 120 are provided.

  Further, the common profile generation unit 41 of the server device 4 generates a common profile for the same situation from all the user data registered in the content selection behavior database 22 and registers it in the common profile database 42. It has become. Here, the common profile 421 registered in the common profile database 42 has the same configuration as the situation-specific profile 121 shown in FIG. 2, and all users are grouped into a plurality of situations. The score value (weight) for content viewing is calculated. Similarly to the situation-specific profile 121, score values are provided for each genre classified into categories (sports, news, drama, etc.) and words (for example, Matsui, Iraq, Diet, etc.).

  Upon receiving the user status from the user status acquisition unit 13, the content priority acquisition unit 31 receives a user profile 121 by status that is a user profile that matches or is closest to the status from the profile database 120, and a long-term user profile unique to the user. The user recommendation 122 is acquired, the content recommendation obtained from the content acquisition unit 12 and the common profile 421 acquired from the common profile database 42 are calculated and transmitted to the recommendation result generation unit 15. It has become.

Here, a method for calculating the recommendation level of content will be specifically described. User i, content (metadata) j, situation k, number of genres m, score value of situation-specific user profile 121 is P (i, k), score value of long-term user profile 122 is L (i), common profile When the score value of 421 is expressed as X (k) and the score value of the content j is expressed as C (j),
P (i, k) = (P _ik1 , P _ik2 , ..., P _ikt , ..., P _ikm )
L (i) = (L _i1 , L _i2 , ..., L _it , ..., L _im )
_{X (k) = (X k1} , X k2, ..., X kt, ..., X km)
C (j) = ( _Cj1 , _Cj2 ..., _Cjt , ..., _Cjm )
It becomes. However, P _ikt , L _it , and X _kt are values in a range of [0, 1] representing weights (score values) regarding the genre t of the user profile 121, the long-term user profile 122, and the common profile 421. _Yes , C _jt is any value of {0, 1} indicating whether the content C (j) is classified into the genre t.

  As an example of calculation of the content recommendation score, a method of expressing the score by the sum of the situation-specific user profile 121, the long-term user profile 122, and the common profile 421 will be described.

In this case, for the content j of the user i in the situation k, the score value S (i, j, k) of the content recommendation degree is represented by the following expression.

However, α _i represents the degree of contribution of the user profile 121 according to the situation of the user i (an index indicating how much importance is placed on the situation-specific profile 121 or the long-term user profile 122 and is set in advance). Is a value in a range of [0, 1] that is sequentially updated by content selection, and β _i represents the contribution degree of the common profile 421 of the user i (the common profile 421, the situation-specific profile 121 and the long-term profile). This is an index indicating how much importance is given to the user profile 122, and is a value in the range of [0, 1] set in advance). Further, the score value S (i, j, k) of the content recommendation level is calculated for all genres t where C _jt = 1 holds.

  When this method is used, the score value S (i, j, k) of the content recommendation level is such that the content that matches the user preference represented by the situation-specific user profile 121 or the long-term user profile 122 or the common profile 421. High score.

  As in the first embodiment, the score value S (i, j, k) of the content recommendation degree is expressed by a method of expressing the score by the product of the user profile 121 by situation, the long-term user profile 122, and the common profile 421. ) May be calculated. When this method is used, the score value S (i, j, k) of the content recommendation degree is a score of content that has a common user preference among the user profile 121 and the long-term user profile 122 according to the situation and the common profile 421. Becomes higher.

  Further, the method for obtaining the score value of the content recommendation level is not limited to the above-described score calculation, and other methods for obtaining the score are also possible. In addition, other information can be used for the data items included in the user profile, and the score may be obtained using the information.

  The client terminal 3 and the server device 4 are electronic devices having a main storage device (memory) including a central processing unit (CPU) having at least a calculation function and a control function, and a RAM having a function of storing programs and data. It consists of a typical device. Each device may include an auxiliary storage device such as a hard disk in addition to the main storage device.

  Among these, the content priority acquisition unit 31 of the client terminal 3 and the common profile generation unit 41 of the server device 4 are nothing but a concrete example of the arithmetic control function by the CPU. The common profile database 42 of the server device 4 has a function of a main storage device.

  A program for executing various processes according to the present embodiment is stored in the main storage device or the hard disk described above. The program can be recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, or a DVD-ROM, or can be distributed via a communication network.

  Next, the operation of the content recommendation system 200 according to the present embodiment will be described using FIG. 7 and FIG. Here, FIG. 7 is a flowchart showing the content recommendation operation in the content recommendation system 200, and FIG. 8 is a flowchart showing the operation in which the content selection behavior of the user is reflected in the profile database 120 and the common profile database 42. It is.

  First, the content recommendation operation when the user accesses the content recommendation system 200 will be described. The access detection unit 11 of the client terminal 3 detects user access and transmits the user ID to the content acquisition unit 12 and the user status acquisition unit 13 (step S310).

  Next, the user status acquisition unit 13 acquires the user status (for example, time information) and transmits it to the content priority acquisition unit 31 (step S320). Further, the content acquisition unit 12 requests content data (content metadata) from the metadata distribution unit 21, and transmits the returned content data to the content priority acquisition unit 31 (step S330).

  Next, the content priority acquisition unit 31 acquires the corresponding situation-specific user profile 121 and long-term user profile 122 from the profile database 120 based on the acquired user ID and user status, and further from the common profile database 42. The common profile 421 is acquired, the score value of content recommendation level is calculated, and transmitted to the recommendation result generation unit 15 (step S340). The recommendation result generation unit 15 rearranges the content in order of score based on the calculated score value of the recommendation level of each content, and transmits it to the recommendation result display unit 16 (step S350).

  Next, the recommendation result display unit 16 displays a list of recommended contents as a recommendation result, and transmits the content selected by the user to the content display unit 17 and the content selection action acquisition unit 18 (step S360).

  Next, the content selection behavior acquisition unit 18 transmits the acquired content selection behavior to the content selection behavior registration unit 19, and the content selection registration unit 19 registers the content selection behavior in the content selection behavior database 22 (step S370). ).

  Next, an operation in which the user's content selection behavior is reflected in the profile database 120 and the common profile database 42 will be described. The content selection database 22 periodically transmits the user's content selection behavior to the profile generation unit 110 of the client terminal 3 according to the user ID from the registered data (step S410).

  The profile generation unit 110 generates a long-term user profile 122 that reflects all user content selection behaviors and does not depend on the situation, and registers it in the profile database 120 (step S420).

  Further, the profile generation unit 110 classifies the user's content selection behavior by situation, generates a situation-specific user profile 121 reflecting the content selection behavior for each situation, and registers the profile in the profile database 120 (step S430).

  Next, the content selection behavior of all users is transmitted from the data registered in the content selection behavior database 22 to the common profile generation unit 110 (step S440).

  The common profile generation unit 110 generates a common profile 421 for all users according to circumstances from the content selection behavior of all users and registers it in the common profile database 42 (step S450).

  Therefore, according to the present embodiment, the user's situation is acquired, and the content of the user profile based on the user's situation is added to the content recommendation so that the user is not burdened. It is possible to recommend the optimum content according to the user's situation.

  This is because, for example, a certain user seems to like a drama from the past viewing history, but does not watch a certain drama program, but among other users, a person who is watching the drama program. Since there are many, the drama program is recommended and presented.

<Third Embodiment>
FIG. 9 is a schematic configuration diagram of a content recommendation system 300 according to the third embodiment of the present invention. A content recommendation system 300 illustrated in FIG. 9 includes a client terminal 5 used by a user and a server device 6 connected to the client terminal 5 via a computer network. In such a system configuration, the server device 6 is a content distribution device that distributes broadcast and VOD type content distribution services represented by digital broadcasting to the client terminal 5, and the client terminal 5 The content distribution service can be enjoyed. In addition, the client terminal 5 uses the content metadata distributed from the server device 6 to present a recommendation of the content that the user will view to the user in consideration of the user's preference according to the user's situation. It is a content recommendation device. Furthermore, in the third embodiment, content recommendation is performed in consideration of a user situation that affects content selection. In the present embodiment, only configurations and functions different from those in the above-described embodiment will be described, and with regard to other configurations and functions, the same portions are denoted by the same reference numerals, and description thereof will be omitted.

  More specifically, the client terminal 5 includes an access detection unit 11, a content acquisition unit 12, a user situation acquisition unit 13, a user situation database 52, a user situation determination unit 53, a content priority acquisition unit 14, a recommendation result generation unit 15, and a recommendation. The result display unit 16, the content display unit 17, the content selection action acquisition unit 18, the content selection action registration unit 51, the profile generation unit 110, and the profile database 120 are configured.

  In addition to the user's content selection behavior sent from the content selection behavior acquisition unit 18, the content selection behavior registration unit 51 includes information indicating the user's situation at the time of content selection (for example, time information from a clock, and from a sensor). From the user position information, clothes, and PIM (Personal Information Manager) information, information indicating a plurality of situations such as future schedules is acquired and registered in the content selection behavior database 61. Therefore, the content selection behavior database 61 stores the user's content selection history together with information indicating the user's situation.

  Then, the user status extraction unit 62 of the server device 6 extracts the user status that seems to have a high correlation with content selection based on the information in the content selection behavior database 61 (statistics based on the accumulated data). For example, in the case of a normal morning situation, when a viewing history is obtained that a news program is being watched, the basis for selecting a news program is “morning”. In the rare case of being in the morning but sitting on the couch, if you have a history of watching a baseball game, the basis for selecting a baseball game is: It can be seen that both “morning” and “sitting on the sofa” are registered in the user situation database 52 on the client terminal 5. It should be noted that the extracted user situation may be composed of a set of a plurality of situations, such as “sitting on the sofa in the“ morning ”” as well as one situation.

  The user situation database 52 stores the user situation extracted by the user situation extraction unit 62 and considered to have a high correlation with content selection.

  The user status determination unit 53 receives access from the user status acquisition unit 13, and uses information obtained from the clock, sensor, PIM information, and the like to indicate information indicating the user status (for example, time information from the clock). , From the sensor, information indicating a plurality of situations such as the user's position information, clothes, and PIM information, such as future schedules), and the situation stored in the user situation database 52 matches or is closest. Is returned to the user status acquisition unit 14.

  The client terminal 5 and the server device 6 are electronic devices having a main storage device (memory) including a central processing unit (CPU) having at least a calculation function and a control function, and a RAM having a function of storing programs and data. It consists of a typical device. Each device may include an auxiliary storage device such as a hard disk in addition to the main storage device.

  Among these, the content selection action registration unit 51 and the user situation determination unit 53 of the client terminal 5 and the user situation extraction unit 62 of the server device 6 are nothing but a concrete example of the arithmetic control function by the CPU. Further, the user situation database 52 of the client terminal 5 and the content selection behavior database 61 of the server device 6 have functions of a main storage device.

  A program for executing various processes according to the present embodiment is stored in the main storage device or the hard disk described above. The program can be recorded on a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, or a DVD-ROM, or can be distributed via a communication network.

  Next, the operation of the content recommendation system 300 according to the present embodiment will be described using FIG. 10 and FIG. Here, FIG. 10 is a flowchart showing the content recommendation operation in the content recommendation system 300, and FIG. 11 is a flowchart for explaining the operation in which the content selection behavior of the user is reflected in the profile database 120 and the user situation database 52. FIG.

  First, the content recommendation operation when the user accesses the content recommendation system 300 will be described. The access detection unit 11 of the client terminal 5 detects the user's access and transmits the user ID to the content acquisition unit 12 and the user status acquisition unit 13 (step S510).

  Next, when the user status acquisition unit 13 accesses the user status determination unit 53 and acquires information indicating the user status obtained from the clock, sensor, and PIM information, the status stored in the user status database 52 Among them, the same or closest situation is acquired and transmitted to the content priority acquisition unit 14 (step S520). In addition, the content acquisition unit 12 requests content data (content metadata) from the metadata distribution unit 21, and transmits the returned content data to the content priority acquisition unit 14 (step S530).

  Next, the content priority acquisition unit 14 acquires the corresponding user profile 211 and long-term user profile 212 from the profile database 120 based on the acquired user ID and user status, and the score value of the content recommendation level Is transmitted to the recommendation result generation unit 15 (step S540). The recommendation result generation unit 15 rearranges the contents in order of score based on the calculated score value of the recommendation level of each content, and transmits it to the recommendation result display unit 16 (step S550).

  Next, the recommendation result display unit 16 displays a list of recommended contents as a recommendation result, and transmits the content selected by the user to the content display unit 17 and the content selection action acquisition unit 18 (step S560).

  Next, the content selection behavior acquisition unit 18 transmits the content selection behavior to the content selection behavior registration unit 51, and the content selection registration unit 51 obtains the content selection behavior from the clock, sensor, PIM information, and the like. Is registered in the content selection behavior database 61 together with information indicating the status of the current (step S570).

  Next, an operation in which a user's content selection behavior is reflected in the profile database 120 and the user situation database 52 will be described. The content selection database 61 periodically transmits the user's content selection behavior from the registered data to the profile generation unit 110 and the user situation extraction unit 62 of the client terminal 5 (step S610).

  The user situation extraction unit 62 extracts a user situation that seems to have a high correlation with content selection, and transmits the user situation to the user situation database 52 (step S620). Thereby, the user situation database 52 stores the user situation that seems to have a high correlation with the content selection (step S630).

  The profile generation unit 110 generates a long-term user profile 122 that reflects all content selection behaviors of the user and does not depend on the situation, and registers it in the profile database 120 (step S640).

  Next, the profile generation unit 110 classifies the user's content selection behavior by situation, generates a situation-specific user profile 121 that reflects the content selection behavior for each situation, and registers it in the profile database 120 (step S650).

  Therefore, according to the present embodiment, the user's situation is acquired, and content recommendation is performed according to the user profile based on the user situation that matches or is close to the user's situation and affects the content selection. It is possible to recommend the optimum content according to the user's situation without burdening the user.

  For example, specifically, according to PIM information, a user seems to go to a concert of an artist next week, and there is a promotion program for this artist, so that the promotion program is recommended and presented. is there. According to the sensor information, since the user is relaxing on the sofa in the morning, it is determined that there is time, and the user recommends and presents a favorite drama or baseball game program.

  In the third embodiment, the content selection action is registered in the content selection database 61 of the server device 6 and then reflected in the profile database 120 and the user situation database 52 on the client terminal 5. The content selection action may be processed only on the client side without being transmitted to the server device 6 and reflected in the profile database 120 and the user situation database 52 as needed.

  While the embodiments of the present invention have been described above, various modifications and changes can be made to the embodiments of the present invention without departing from the spirit of the present invention. For example, in the above-described embodiment, a case where one user uses one client terminal has been described. However, the present invention is not limited to this, and one client terminal can be used by a plurality of users. Of course, you can share it. However, in this case, the profile database 120 and the user situation database 52 are managed for each user, and the corresponding user data is acquired by the user ID logged into the client terminal. Is.

It is a schematic block diagram of the content recommendation system which concerns on the 1st Embodiment of this invention. It is a figure explaining the structural example of the user file classified by situation. It is a figure explaining the structural example of a long-term user file. It is a flowchart figure which shows the operation | movement of the content recommendation in the content recommendation system which concerns on the 1st Embodiment of this invention. It is a flowchart figure which shows the operation | movement in which the content selection action of a user is reflected in a profile database in the content recommendation system which concerns on the 1st Embodiment of this invention. It is a schematic block diagram of the content recommendation system which concerns on the 2nd Embodiment of this invention. It is a flowchart figure which shows the operation | movement of the content recommendation in the content recommendation system which concerns on the 2nd Embodiment of this invention. It is a flowchart figure which shows the operation | movement in which the content selection action of a user is reflected in a profile database and a common profile database in the content recommendation system which concerns on the 2nd Embodiment of this invention. It is a schematic block diagram of the content recommendation system which concerns on the 3rd Embodiment of this invention. It is a flowchart figure which shows the operation | movement of content recommendation in the content recommendation system which concerns on the 3rd Embodiment of this invention. It is a flowchart figure explaining the operation | movement by which a user's content selection action is reflected in a profile database and a user condition database in the content recommendation system which concerns on the 3rd Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,3,5 Client terminal 2,4,6 Server apparatus 11 Access detection part 12 Content acquisition part 13 User condition acquisition part 14,31 Content priority acquisition part 15 Recommendation result generation part 16 Recommendation result display part 17 Content display part 18 Content selection action acquisition unit 19, 51 Content selection action registration unit 21 Metadata distribution unit 22, 61 Content selection action database 23 Content distribution unit 41 Common profile generation unit 42 Common profile database 52 User situation database 53 User situation determination unit 62 User situation Extraction unit 100, 200, 300 Content recommendation system 110 Profile generation unit 120 Profile database 121 User profile by situation 122 Long-term user profile

Claims (18)

A content distribution apparatus that distributes a plurality of contents can communicate with each other via a computer network, and a user will watch using metadata that describes the contents acquired from the content distribution apparatus. A content recommendation device for recommending content,
A situation-specific user profile storage means for storing the user's content viewing history as first preference information for each viewing situation;
Long-term user profile storage means for storing the user's content viewing history as second preference information for each user regardless of viewing status;
Viewing status acquisition means for acquiring the viewing status of the user when the user accesses the content recommendation device;
Metadata acquisition means for acquiring the metadata of the plurality of contents;
The first preference information corresponding to the viewing situation that matches or is similar to the viewing situation acquired by the viewing situation acquisition unit for each of the acquired metadata of the plurality of contents is the user profile storage unit for each situation And the relationship between the acquired first preference information and the second preference information and the content of the content obtained from the metadata, and the respective contribution degrees of the first and second preference information A recommendation degree calculating means for calculating a score related to content recommendation based on
Sorting the scores for each content, extracting the content to be recommended, and displaying the recommended result to the user;
A content recommendation device comprising:   The content selection information selected by the user from among the content to be recommended displayed to the user is the first preference information of the user profile storage means by situation and the long-term user profile storage means. The content recommendation device according to claim 1, wherein the content recommendation device is reflected in the second preference information. A common profile acquisition unit that acquires, from the content distribution device, third preference information obtained by calculating the content viewing history of all the users for each viewing situation;
The content recommendation device according to claim 1, wherein the recommendation degree calculation unit further takes into account the third preference information and a contribution degree of the third preference information. Viewing status storage means for storing a viewing status that has influenced the content selection when the user selects content from the content to be recommended displayed to the user;
The viewing status acquisition unit acquires one or a plurality of viewing statuses, acquires a viewing status that matches or is similar to the acquired viewing status from the viewing status storage unit,
The content recommendation device according to any one of claims 1 to 3, wherein the recommendation degree calculation unit acquires the first preference information according to a viewing situation acquired from the viewing situation storage unit.   The preference information has a predetermined score value having a higher score as the user's preference is higher for each genre classified into a content category and a word indicating content content, and the metadata includes the genre 5. The content recommendation device according to claim 1, wherein the content recommendation device has a predetermined score value when the content corresponds to the genre.   The content recommendation apparatus according to claim 1, wherein the viewing status includes at least one of time information and position information, schedule information, and clothes information of the user. A content recommendation apparatus capable of communicating with each other via a computer network and a content distribution apparatus that distributes a plurality of contents is viewed by a user using metadata that describes the contents of the content acquired from the content distribution apparatus A content recommendation method for recommending content that would be
A user profile storage step for each situation in which the content viewing history of the user is stored in the storage unit as first preference information for each viewing situation;
A long-term user profile storage step of storing the content viewing history of the user in the storage unit as second preference information for each user regardless of viewing status;
A viewing status acquisition step of acquiring the viewing status of the user when the user accesses the content recommendation device;
A metadata acquisition step of acquiring the metadata of the plurality of contents;
The first preference information corresponding to the viewing situation that matches or is similar to the viewing situation obtained in the viewing situation obtaining step is obtained from the storage unit for each of the obtained metadata of the plurality of contents, and obtained. Content based on the relationship between the first preference information and the second preference information and the content of the content obtained from the metadata, and the respective contributions of the first and second preference information. A recommendation degree calculating step for calculating a score related to the recommendation;
Sorting the scores for each content, extracting the content to be recommended, and displaying a recommendation result to be displayed to the user;
The content recommendation method characterized by having.   Content selection information selected by the user from contents to be recommended displayed to the user is reflected in the first preference information and the second preference information of the storage unit. The content recommendation method according to claim 7. A common profile acquisition step of acquiring third preference information obtained by calculating the content viewing history of all the users for each viewing situation from the content distribution device;
The content recommendation method according to claim 7, wherein the recommendation degree calculation step further takes into account the third preference information and a contribution degree of the third preference information. A viewing status storage step of storing, in the viewing status storage unit, a viewing status that has influenced content selection when the user selects content from among the content to be recommended displayed to the user;
The viewing status acquisition step acquires one or a plurality of viewing statuses, acquires a viewing status that matches or is similar to the acquired viewing status from the viewing status storage unit,
The content recommendation method according to any one of claims 7 to 9, wherein the recommendation degree calculation step acquires the first preference information according to a viewing situation obtained from the viewing situation storage unit.   The preference information has a predetermined score value having a higher score as the user's preference is higher for each genre classified into a content category and a word indicating content content, and the metadata includes the genre The content recommendation method according to any one of claims 7 to 10, wherein each content has a predetermined score value when the content corresponds to the genre.   The content recommendation method according to any one of claims 7 to 11, wherein the viewing status includes at least one of time information and position information, schedule information, and clothes information of the user. A computer that can communicate with a content distribution device that distributes a plurality of contents via a computer network uses a metadata that describes the content acquired from the content distribution device to be viewed by a user. A content recommendation program for recommending wax content,
A user profile storage step for each situation in which the content viewing history of the user is stored in the storage unit as first preference information for each viewing situation;
A long-term user profile storage step of storing the content viewing history of the user in the storage unit as second preference information for each user regardless of viewing status;
A viewing status acquisition step of acquiring the viewing status of the user when the user accesses the content recommendation device;
A metadata acquisition step of acquiring the metadata of the plurality of contents;
The first preference information corresponding to the viewing situation that matches or is similar to the viewing situation obtained in the viewing situation obtaining step is obtained from the storage unit for each of the obtained metadata of the plurality of contents, and obtained. Content based on the relationship between the first preference information and the second preference information and the content of the content obtained from the metadata, and the respective contributions of the first and second preference information. A recommendation degree calculating step for calculating a score related to the recommendation;
Sorting the scores for each content, extracting the content to be recommended, and displaying a recommendation result to be displayed to the user;
Is executed by the computer.   Content selection information selected by the user from contents to be recommended displayed to the user is reflected in the first preference information and the second preference information of the storage unit. The content recommendation program according to claim 13. Causing the computer to execute a common profile acquisition step of acquiring third preference information obtained by calculating the content viewing history of all the users for each viewing situation from the content distribution device;
The content recommendation program according to claim 13 or 14, wherein the recommendation degree calculation step further takes into account the third preference information and a contribution degree of the third preference information. A viewing state storage step of storing in the viewing state storage unit the viewing state that has influenced the content selection when the user selects content from the content to be recommended displayed to the user. Let it run
The viewing status acquisition step acquires one or a plurality of viewing statuses, acquires a viewing status that matches or is similar to the acquired viewing status from the viewing status storage unit,
The content recommendation program according to any one of claims 13 to 15, wherein the recommendation degree calculation step acquires the first preference information according to a viewing situation obtained from the viewing situation storage unit.   The preference information has a predetermined score value having a higher score as the user's preference is higher for each genre classified into a content category and a word indicating content content, and the metadata includes the genre The content recommendation program according to claim 13, wherein the content recommendation program has a predetermined score value when the content corresponds to the genre.   The content recommendation program according to any one of claims 13 to 17, wherein the viewing situation includes at least one of time information and position information, schedule information, and clothes information of the user.

Images