JP5232495B2 - Content distribution control device and computer program - Google Patents

Description

  The present invention relates to a content distribution control device and a computer program.

Conventionally, as a content distribution system, for example, those described in Patent Documents 1 and 2 are known. In the prior art 1 described in Patent Document 1, a plurality of content storage units are provided, and content having a high user preference is arranged in the content storage unit closest to the user terminal. In Prior Art 2 described in Patent Document 2, a content reception / playback device receives and plays back content based on conditions such as user preference, user-side communication environment, content genre, and the like from content that can be distributed by a content server. The content to be performed is selected and the playback order of the content is determined and transmitted to the content server.
Japanese Patent Laid-Open No. 2003-30087 Japanese Patent Laid-Open No. 2005-84752

However, the conventional technology described above has the following problems.
According to the prior art 1, by providing a plurality of content storage means, it is possible to reduce the server load at the time of content distribution, but there is a problem in realization that it increases the cost of facilities.
According to the prior art 2, content distribution according to the conditions of individual content reception / playback apparatuses is possible. However, in order to meet individual requests from a large number of content reception / playback apparatuses, the required processing capacity of the content server is great. There is concern about becoming a thing.

Generally, there is a limit to the processing capability that can be provided in the server, and it can be easily imagined that there is an upper limit of resources that can be used for content distribution. On the other hand, it is desirable for individual users to distribute content that suits their taste as much as possible. Here, a method of limiting the number of distribution destination users for each content can be considered, but there are the following problems (1) and (2).
(1) By setting a uniform upper limit on the number of content distributions, the load on the distribution server can be kept constant. However, since the degree of popularity (number of users who want the content) differs for each content, User satisfaction is reduced.
(2) The user's satisfaction can be improved by calculating an index such as the similarity between the content and the user's preference and then distributing it only to users exceeding a certain value. The ability may be exceeded.
For this reason, it is an issue to perform content distribution control so as to distribute to each user content that matches the preferences of individual users as much as possible within the range of content distribution capability.

  The present invention has been made in consideration of such circumstances, and the purpose of the present invention is to distribute content so that each user can distribute content that suits each user's preference as much as possible within the range of content distribution capability. It is an object of the present invention to provide a content distribution control device and a computer program capable of controlling the content.

  In order to solve the above problems, the content distribution control device according to the present invention distributes a surplus of resources available for content distribution to a new distribution target content according to the popularity of the content, and distributes surplus resources. Content distribution number calculating means for calculating the number of distributions for each new distribution target content based on the result, and the distribution target based on the similarity between the user preference and the distribution target content grasped from the user's content browsing history And a first content distribution destination selection unit that selects a number of distribution destination users according to the number of content distributions.

  In the content distribution control apparatus according to the present invention, the popularity of the new distribution target content is increased or decreased according to the tendency of the viewer rate calculated from the user browsing history related to the content similar to the new distribution target content. It is characterized by being.

  In the content distribution control device according to the present invention, the user preference is calculated as a vector that combines the feature vectors of the content browsed by the user, and the first content distribution destination selecting means includes the calculation result vector and The similarity is calculated with the feature vector of the new distribution target content.

  In the content distribution control apparatus according to the present invention, the first content distribution destination selection means selects a distribution destination user based on a user participation rate calculated from a content distribution history to the user and a user's content browsing history. It is characterized by reflecting.

  In the content distribution control device according to the present invention, a non-distributed user information storage unit for storing information on a non-distributed user that is lost from selection of a distribution destination user by the first content distribution destination selection unit, and a new distribution target Second content delivery destination selection means for selecting a delivery destination user within the range of the surplus resources for undelivered users when there is no content and there is a surplus of resources available for content delivery; , Provided.

  In the content distribution control device according to the present invention, the content distribution number calculation means preferentially distributes the surplus of resources available for content distribution to the priority content indicated by the priority content information, and then the remaining content The surplus resources are distributed to non-priority contents according to their popularity.

  In the content distribution control device according to the present invention, the first content distribution destination selection means selects the priority user indicated by the priority user information as a distribution destination preferentially, and then distributes the remaining number of distribution destinations. It is characterized by selecting a user.

The computer program according to the present invention distributes a surplus of resources that can be used for content distribution to a new distribution target content according to the popularity of the content, and sets each new distribution target content based on the surplus resource distribution result. The number of distribution destinations corresponding to the number of distribution of the distribution target content based on the content distribution number calculation function for calculating the distribution number of the content and the similarity between the user preference and the distribution target content ascertained from the user's content browsing history A first content distribution destination selection function for selecting a user is realized in a computer.
As a result, the above-described content distribution control apparatus can be realized using a computer.

  According to the present invention, it is possible to control the distribution of content so that each user can distribute content that matches the preferences of individual users as much as possible within the range of content distribution capability.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic configuration diagram showing the overall configuration of a content distribution system according to an embodiment of the present invention. In FIG. 1, a distribution server 1 receives new content from a web site 2 via the Internet 3. In FIG. 1, the website 2 connected to the Internet 3 is illustrated. However, the distribution server 1 receives new content from the website 2 connected to a local area network (LAN) such as a company via the LAN. It may be received.

Examples of the website 2 include a blog site, a news site, an image posting site, a music posting site, and the like. Newly arrived content is newly registered content on the website 2. For example, articles newly posted on the blog site, articles newly posted on the news site, images newly posted on the image posting site, music posting sites Newly posted songs and so on. Content is composed of data such as text, images, and audio.
In the present embodiment, the content includes text data as an example.

  The distribution server 1 distributes newly arrived content to the user terminal 4 via the communication network 5. The communication network 5 is used for content distribution. The communication network 5 includes a wired network or a wireless network, and may include the Internet 3.

  The user terminal 4 is configured to receive content via the communication network 5 and allow the user to browse the received content. Further, the user terminal 4 is configured to transmit information on the date and time when the user browses the received content together with the content identification information (content ID) of the browsed content to the distribution server 1 via the communication network 5. The browsed content ID and the browsing date information are information (participation information) indicating that the user has participated in browsing newly arrived content.

  The user terminal 4 may be a device that can be used while moving (for example, a portable terminal device), or a device that is used fixedly (for example, a personal computer or a set-top box (for example, cable television broadcasting). It may be a stationary terminal device such as STB for reception). Examples of the portable terminal device include a mobile communication network terminal device provided by a mobile communication carrier (for example, a mobile phone having a data communication function, a data communication terminal device, etc.), a PDA (Personal Digital Assistants: personal). Information equipment). In the case of a PDA, a communication unit may be built in or a communication unit may be connected from the outside.

  FIG. 2 is a block diagram illustrating a configuration of the distribution server 1 according to the present embodiment. In FIG. 2, the distribution server 1 includes a new content receiver 11, a new content storage 12, a feature vector calculator 13, a base category database (base category DB) 14, a participation information receiver 15, a participation information storage 16, and a distribution. A control unit 17 and a distribution unit 18 are included.

  The new arrival content receiving unit 11 receives new arrival content from the website 2 via the Internet 3. The new arrival content receiving unit 11 stores the received content in the new arrival content storage unit 12 together with information on the date and time when the new arrival content is received. FIG. 3 shows a configuration example of the newly arrived content storage unit 12. The new arrival content storage unit 12 stores content data, reception date / time, feature vector, and priority flag in association with the content ID for each content. The content data includes content main body data and attribute data (such as content URL (Uniform Resource Locator) and base category information). The content for which the priority flag is on is a priority content that should be given priority for delivery (for example, an article with high urgency such as an earthquake early warning). Content whose priority flag is off is non-priority content. The base category information is set from the feature vector calculation unit 13. The feature vector is calculated and set by the feature vector calculation unit 13.

  The feature vector calculation unit 13 receives a copy of the new arrival content from the new arrival content reception unit 11. At this time, the newly arrived content receiving unit 11 may copy only the data used for the feature vector calculation process. In this embodiment, since only text data is used in the feature vector calculation process, the newly arrived content receiving unit 11 copies only the text data included in the newly arrived content and sends it to the feature vector calculating unit 13. The feature vector calculation unit 13 calculates a feature vector of the newly arrived content from a copy (here, text data) of the newly arrived content received from the newly arrived content receiving unit 11.

  The feature vector calculation unit 13 performs feature vector calculation processing using the base category DB 14. The base category DB 14 stores a feature word list for each type of information (base category) defined in advance. The feature word list is generated based on the learning data for each base category. Examples of the base category include entertainment, sports, gourmet, politics, and economy.

  Here, an example of a method for generating a feature word list will be described focusing on a base category. First, words such as nouns and sub-variables are extracted from a plurality of learning data (text data) classified into the base category using a tool such as morphological analysis. Next, an importance value (for example, TF / IDF value) based on the appearance probability of a general word is calculated for the extracted word. Next, the importance value and extracted word pairs are stored in the feature word list. This generation process is performed for each base category, a feature word list for each base category is generated, and stored in the base category DB 14.

  The base category DB 14 can be constructed independently of other processes. Further, it is preferable to periodically update the feature word list of each base category with new learning data.

Next, with reference to FIG. 4, an operation in which the feature vector calculation unit 13 calculates a feature vector of new arrival content will be described. FIG. 4 is a flowchart showing the procedure of the feature vector calculation process according to the present embodiment.
In FIG. 4, in step S1, a feature word list of newly arrived content is created. In this feature word list creation process, first, words such as nouns and sub-variables are extracted from text data of newly arrived content received from the newly arrived content receiving unit 11 using a tool such as morphological analysis. Next, the number of times that the extracted word appears in the text data is counted. Next, the value of the number of appearances and the set of extracted words are stored in the feature word list.

  In step S2, the degree of association with newly arrived content is calculated for each base category. In this relevance calculation process, first, the inner product of the feature word list of the new content and the feature word list of the base category is calculated for each base category. In this inner product calculation, all the extracted words included in common in both feature word lists are multiplied by the appearance frequency value and the importance value, and the products are summed. This total value is the value of the relevance of newly arrived content with respect to the base category. This process is performed for each base category, and the relevance value of newly arrived content for each base category is calculated.

  In step S3, a set of relevance values for each base category calculated in step S2 is output as a feature vector of newly arrived content. This feature vector is stored in the newly arrived content storage unit 12 as a feature vector of the newly arrived content and is sent to the distribution control unit 17. Further, the feature vector calculation unit 13 stores the base category information of the new arrival content in the new arrival content accumulation unit 12. The base category information indicates a base category whose degree of association with newly arrived content is a predetermined value or more.

Returning to FIG.
The participation information receiving unit 15 receives participation information (viewed content ID and browsing date information) from the user terminal 4 via the communication network 5. The participation information receiving unit 15 stores the received participation information in the participation information accumulating unit 16. FIG. 5 shows a configuration example of the participation information storage unit 16. The participation information storage unit 16 stores user information, content distribution history, and content browsing history in association with the user ID for each user.

  The user information includes a user address and a priority flag. The user address is a communication address of the user terminal 4 and is used at the time of content distribution. A user whose priority flag is on is a priority user who should give priority to content distribution (for example, a user who subscribes to a paid distribution service). A user whose priority flag is off is a non-priority user. The user information is set in advance and can be arbitrarily updated.

  The content distribution history includes the ID of content distributed to the user terminal 4, information on the distribution date and time, and information on the number of distributed content. The content distribution history is set from the distribution control unit 17. The distribution control unit 17 stores the ID of the content already distributed to the user terminal 4 and information on the distribution date and time in the participation information storage unit 16. At this time, the distribution control unit 17 updates the number of distributed contents by increasing the number of distributed contents by the number of distributed contents.

  The content browsing history includes the ID of content that has been browsed on the user terminal 4, information on the browsing date and time, and information on the number of browsed content. The content browsing history is set from the participation information receiving unit 15. The participation information receiving unit 15 stores the participation information (viewed content ID and browsing date information) received from the user terminal 4 in the participation information accumulating unit 16. At this time, the participation information receiving unit 15 increases the number of browsed contents by the number of received browsed content IDs, and updates the information on the number of browsed contents.

  The distribution control unit 17 is configured to control content distribution in order to improve user satisfaction within the range of content distribution capability. The distribution control unit 17 determines a distribution destination user of the newly arrived content and outputs the information to the distribution unit 18. Details of the distribution control unit 17 will be described later.

  The distribution unit 18 distributes each newly arrived content to the user terminal 4 of each distribution destination user based on the distribution destination user information received from the distribution control unit 17. The distribution unit 18 acquires the user address of each distribution destination user from the participation information storage unit 16. In addition, the distribution unit 18 acquires each new content from the new content storage unit 12.

  FIG. 6 is a block diagram illustrating a configuration of the distribution control unit 17 according to the present embodiment. The distribution control unit 17 is configured to control content distribution in order to improve user satisfaction within the range of content distribution capability. In FIG. 6, the distribution control unit 17 includes a content maximum distribution number calculation unit 21, a content distribution number calculation unit 22, a first content distribution destination selection unit 23, a second content distribution destination selection unit 24, a priority content information storage unit 25, A priority user information storage unit 26 and an undistributed user information storage unit 27 are included.

  The distribution control unit 17 executes content distribution control processing at a constant cycle. When the distribution control unit 17 receives the feature vector of the newly arrived content from the feature vector calculation unit 13, the distribution control unit 17 holds it until the content distribution control process is executed. Therefore, when the content distribution control process is executed, it may happen that feature vectors related to a plurality of newly arrived contents are held. When it is time to start the content distribution control process, the distribution control unit 17 executes the content distribution control process using the feature vectors related to all the new arrival contents that are held. In addition, all newly arrived content related to the retained feature vector is a content distribution control processing target.

  Hereinafter, the operation related to the content distribution control process of the distribution control unit 17 will be described in detail with reference to the flowcharts of FIGS.

  First, the operation of the content maximum distribution number calculation unit 21 will be described with reference to FIGS. In FIG. 7, in step S <b> 11, it is determined whether there is priority content among newly arrived content. In this determination process, the priority content information stored in the priority content information storage unit 25 is used. The priority content information includes information (prior content URL or the like) for specifying the priority content. The content maximum distribution number calculation unit 21 obtains the URL of the newly arrived content from the newly arrived content storage unit 12 and collates it with the prioritized content information. Alternatively, the maximum content distribution number calculation unit 21 acquires the priority flag of the new arrival content from the new arrival content storage unit 12, and determines that there is priority content when the priority flag is on. If there is priority content, the process proceeds to step S12. If there is no priority content, the process proceeds to step S14.

  In step S12, a transmission queue is secured for newly arrived content determined to be priority content. The transmission queue is one resource management unit that can be used for content distribution in the distribution unit 18. The transmission queue corresponds to a memory area that holds content waiting to be transmitted. In this embodiment, one transmission queue is assigned to one content distribution. For example, when one content is distributed to two user terminals 4 respectively, two transmission queues are required. That is, one transmission queue is assigned to a set of one content and one distribution destination. Therefore, the maximum content distribution number calculation unit 21 secures transmission queues corresponding to the number of combinations of the priority content number and the user number.

  In step S13, a value obtained by subtracting the number of priority contents from the number of newly arrived contents is set in the variable n.

  In step S14, since it is determined in step S11 that there is no priority content in the newly arrived content, the number of newly arrived content is set to the variable n as it is.

  In step S15, it is determined whether or not the transmission queue is empty. Information on the number of unused transmission queues is received from the distribution unit 18. If there is an unused transmission queue, the process proceeds to step S16, and if there is no unused transmission queue, the process proceeds to step S17.

In step S16, the number of unused transmission queues (the number of empty transmission queues) is set in the variable Q. Also, an initial value “1” is set to the variable i.
On the other hand, in step S17, since there is no vacant transmission queue, the newly arrived content is stored in the spare queue until the next content distribution control process is executed.

Next, in steps S18 to S26 in FIG. 8, the maximum number of distributions is calculated for each of n new arrival contents.
First, in step S18, the i-th new arrival content is set as the processing target content.
In step S19, it is determined whether or not there has been a predetermined number (k times) of similar content in the past. This process will be described. First, the content maximum distribution number calculation unit 21 refers to the base category information stored in the newly arrived content storage unit 12 for the processing target content, and grasps the base category to which the processing target content belongs. Next, another content belonging to the base category is searched from the newly arrived content storage unit 12. Next, the number of contents (similar contents) as a result of the search is counted, and if it is k or more, it is determined that k similar contents existed in the past. If k similar contents exist in the past, the process proceeds to step S20. If not, the process proceeds to step S22.

  In step S20, each viewer rate for the past k similar contents is calculated. This process will be described. First, the maximum content distribution number calculation unit 21 acquires the reception date and time of the similar content of the search result in step S19 from the new content storage unit 12, and selects k similar content from the newest reception date and time. Next, a viewer rate is calculated for each of the k similar contents. The viewer rate is a value obtained by dividing the number of users who have viewed similar content by the number of distribution destination users who have distributed the similar content. The number of users who have viewed similar content is obtained by searching the content browsing history in the participation information storage unit 16. The number of distribution destination users who have distributed similar content is obtained by searching the content distribution history in the participation information storage unit 16.

  In step S21, the tendency (up, down, leveling) of the viewer rate is determined based on each viewer rate for the past k similar contents. About the tendency of a visitor rate, it grasps | ascertains by calculating the inclination between visitor rates based on the reception date of similar content.

  In steps S22, S23, and S24, the maximum number of distributions Pxi of the processing target content is calculated according to the result of step S21. If the viewer rate tends to be flat, the process proceeds to step S22, and the current maximum number of distributions Pxi is maintained. If the viewer rate is increasing, the process proceeds to step S23, and the maximum number of distributions Pxi is increased. If the viewer rate is in a downward trend, the process proceeds to step S24, and the maximum distribution number Pxi is decreased. Note that the initial value of the maximum distribution number Pxi is set in advance. Moreover, as a method of increasing / decreasing the maximum distribution number Pxi, for example, increasing or decreasing a predetermined rate can be mentioned. However, the lower limit value of the maximum delivery number Pxi is 1.

  In step S25, 1 is added to the variable i. In step S26, it is determined whether the variable i is larger than the variable n. If the variable i is greater than the variable n, the processing for all newly arrived content has been completed, so the content maximum distribution number calculation processing is terminated. On the other hand, if the variable i is less than or equal to the variable n, since unprocessed newly arrived content remains, the process returns to step S18, and the next unprocessed newly arrived content is set as a new process target content to calculate the maximum content distribution number. Continue processing.

  The content maximum distribution number calculation unit 21 outputs each maximum distribution number for the n new arrival contents to the content distribution number calculation unit 22.

Next, the operation of the content distribution number calculation unit 22 will be described.
The content distribution number calculation unit 22 distributes an unused transmission queue to n new arrival contents according to each maximum distribution number, and calculates the distribution number for each new arrival content based on the distribution result. This process will be described. First, the content distribution number calculation unit 22 calculates the sum Pa of each maximum distribution number Pxi (where 1 ≦ i ≦ n) for n new arrival contents. Next, the distribution number Pi is calculated for each new content. The distribution number Pi is calculated by the following equation.
Pi = Q × (Pxi ÷ Pa)
However, Q is the value of the variable Q, which is the number of empty transmission queues.

  In this calculation result, the distribution number Pi of a new arrival content corresponds to the distribution number of empty transmission queues for the new arrival content. The distribution of the empty transmission queue is in accordance with the ratio of the maximum number of distributions of the newly arrived content in the sum of the maximum number of distributions of the n new arrival contents. Here, as described above, the maximum number of newly delivered contents is increased or decreased according to the tendency of the viewer rate, and represents the popularity of the newly arrived contents. Therefore, the distribution of the empty transmission queue is performed according to the popularity of newly arrived content. The empty transmission queue corresponds to the surplus of resources that can be used for content distribution. Therefore, the distribution number Pi of newly arrived content is calculated based on the surplus resource distribution result.

  The content distribution number calculation unit 22 outputs each distribution number for the n new arrival contents to the first content distribution destination selection unit 23.

Next, the operation of the first content distribution destination selection unit 23 will be described with reference to FIG.
In FIG. 9, in step S31, the number of users is set in a variable m. Also, an initial value “0” is set in the variable i. The number of users is the number of users registered in the participation information storage unit 16.

  Next, in steps S32 to S42, the number of distribution destination users corresponding to the number of distributions is selected for n new arrival contents.

  First, in step S32, 1 is added to the variable i, and in step S33, the i-th new arrival content is set as the processing target content. In step S34, an initial value “0” is set to the variable j.

  Next, in steps S35 to S39, each of m users is processed.

  First, in step S35, 1 is added to the variable j, and the j-th user is set as a processing target user. In step S36, the similarity S (i, j) between the feature vector Xi of the processing target content and the processing target user's preference vector Mj is calculated. This calculation process will be described. First, the first content distribution destination selection unit 23 calculates a preference vector Mj of the processing target user. In this preference vector calculation process, all the browsed content IDs of the processing target user are acquired from the participation information storage unit 16. Next, the feature vector corresponding to each browsed content ID is acquired from the new arrival content storage unit 12. Next, a vector that combines the acquired feature vectors is calculated. For example, the sum of feature vectors is calculated. Alternatively, the average of feature vectors is calculated. The vector of the calculation result (for example, the sum of the feature vectors, the average, etc.) is the processing target user's preference vector Mj. Next, the first content distribution destination selection unit 23 calculates the similarity S (i, j) between the feature vector Xi of the processing target content and the preference vector Mj of the processing target user. As the similarity S (i, j), for example, a cosine distance, an inner product, an Euclidean distance, or the like can be used.

In step S37, the participation rate Rj of the processing target user is calculated. The participation rate Rj is calculated by the following equation.
Rj = Rcj ÷ Roj
Here, Rcj is the browsing rate for the processing target user. Roj is the number of newly arrived contents after the point in time when the user to be processed browses the contents last. The browsing rate Rcj is a value obtained by dividing the number of browsed contents in the participation information storage unit 16 for the processing target user by the number of distributed contents. The newly arrived content number Roj after the last viewing time is obtained by counting the number of contents corresponding to the reception date and time in the new arrival content storage unit 12 from the latest browsing date and time in the participation information storage unit 16 for the processing target user. Desired.

  In step S38, the similarity S (i, j) is multiplied by the participation rate Rj. This product “S (i, j) × Rj” is held.

  In step S39, it is determined whether the variable j is equal to the variable m. If the variable j is equal to the variable m, the process of steps S35 to S39 is completed for all users, and the process proceeds to step S40. On the other hand, if the variable j is less than the variable m, there remains an unprocessed user, so the process returns to step S35, the next unprocessed user is set as a new process target user, and the processes of steps S35 to S39 are performed. continue.

  Through steps S35 to S39, the value of “S (i, j) × Rj” is calculated for m users.

  In step S40, the distribution ranking of the processing target contents is performed for m users. In this distribution ranking process, first, a priority user is placed at a higher rank. The priority user is indicated by the priority user information stored in the priority user information storage unit 26. Next, the remaining users are arranged in descending order of the value of “S (i, j) × Rj”.

  In step S41, the top Pi persons are selected as distribution destination users of the processing target content according to the distribution order. Here, Pi is the number of distributions of the content to be processed, and is calculated by the content distribution number calculation unit 22 described above. Next, the value of “S (i, j) × Rj” of the remaining users (undistributed users) that has been lost from the selection of the distribution destination user this time is recorded in the undistributed user information storage unit 27 as undistributed user information. To do.

  In step S42, it is determined whether the variable i is equal to the variable n. If the variable i is equal to the variable n, the processing for all the newly arrived content has been completed, so the first content distribution destination selection processing is terminated. On the other hand, if the variable i is less than the variable n, the process returns to step S32 because unprocessed newly arrived content remains, and the next unprocessed newly arrived content is set as a new process target content in step S33. Continue the content distribution destination selection process.

  The first content distribution destination selection unit 23 outputs information on each distribution destination user regarding the n new arrival contents to the distribution unit 18. The distribution unit 18 distributes each newly arrived content to the user terminal 4 of each distribution destination user based on the distribution destination user information. The distribution unit 18 acquires the user address of each distribution destination user from the participation information storage unit 16. In addition, the distribution unit 18 acquires each new content from the new content storage unit 12.

Next, the operation of the second content distribution destination selection unit 24 will be described with reference to FIG.
The second content distribution destination selection unit 24 selects a distribution destination user for undistributed users who have lost the selection of the distribution destination user by the first content distribution destination selection unit. The second content distribution destination selection unit 24 selects a distribution destination user within the range of the empty transmission queue for undistributed users when no new newly received content is received and the transmission queue is empty. Select.

  The second content distribution destination selection unit 24 confirms that a feature vector of newly arrived content has not been newly received from the feature vector calculation unit 13 after selection of a distribution destination user by the first content distribution destination selection unit. Further, the second content distribution destination selection unit 24 receives information on the number of unused transmission queues from the distribution unit 18 and confirms that there is an empty transmission queue. The second content distribution destination selection unit 24 starts the second content distribution destination selection process shown in FIG. 10 when both of them are confirmed.

In FIG. 10, in step S51, the number of empty transmission queues is set in the variable Q.
In step S52, undistributed user information for all (n) newly arrived contents is read from the undistributed user information storage unit 27, and for all (Nu) undistributed users included in all undistributed user information. , “S (i, j) × Rj” (where 1 ≦ i ≦ n, 1 ≦ j ≦ m) is assigned in descending order.

  In step S53, it is determined whether the variable Q is equal to or greater than the total number of undelivered users “Nu”. If the variable Q is greater than or equal to Nu, the process proceeds to step S54, and if the variable Q is less than Nu, the process proceeds to step S55.

  In step S54, since the number of empty transmission queues is sufficient for all the undistributed users, all Nu users of the undistributed users are selected as distribution destinations of the corresponding newly arrived content. Further, all undelivered user information is deleted from the undelivered user information storage unit 27.

  In step S55, the number of empty transmission queues is insufficient for all undistributed users, so the top Q undistributed users are selected as distribution destinations of the corresponding new content according to the distribution order in step S52. Further, the non-delivery user information of the selected user is deleted from the non-delivery user information storage unit 27.

  The second content distribution destination selection unit 24 outputs information of each distribution destination user additionally selected by the second content distribution destination selection process to the distribution unit 18.

  The above is the detailed description of the distribution control unit 17 according to the present embodiment.

  As described above, according to the present embodiment, the free space in the transmission queue corresponding to the surplus of resources available for content distribution is distributed to newly arrived content according to the popularity of the content, and based on the distribution result, The number of distributions for each new content is calculated. Furthermore, the number of distribution destination users corresponding to the number of distribution of the new arrival content is selected based on the similarity between the user preference and the new arrival content grasped from the user's content browsing history.

  This makes it possible to distribute newly arrived content within the range of content distribution capability based on the surplus of resources available for content distribution. Furthermore, content distribution control can be performed by the distribution server side without being influenced by the number of users participating in the content distribution service. In addition, more popular new contents can be distributed. Furthermore, it is possible to distribute newly arrived content to a preferred user based on user preferences. In other words, it is possible to realize that new contents are not distributed to users who are not interested. As a result, effective use of resources is achieved, and it is possible to distribute to each user content that matches the preferences of individual users as much as possible within the range of content distribution capability. Thereby, while controlling the load concerning a delivery server appropriately, the special effect that a user satisfaction can be improved by delivering an appropriate new arrival content with respect to a user is acquired.

  Further, according to the present embodiment, the popularity degree of newly arrived content is increased or decreased according to the tendency of the viewer rate calculated from the user browsing history related to the content similar to the newly arrived content. Thereby, it is possible to adaptively distribute surplus resources according to the user's browsing tendency and calculate the number of distributions of each newly arrived content.

  Further, according to the present embodiment, the user preference is calculated as the sum of the feature vectors of the content viewed by the user. Thereby, an appropriate delivery destination user can be selected by calculating the similarity with the feature vector of newly arrived content.

  Further, according to the present embodiment, the participation rate of the user calculated from the content distribution history to the user and the user's content browsing history is reflected in the selection of the distribution destination user. Thereby, effective delivery can be increased and it can contribute to effective use of resources.

  In addition, according to the present embodiment, the information on the non-delivery user that has been lost from the selection of the delivery destination user by the first content delivery destination selection process is recorded, there is no new newly arrived content, and the transmission queue is empty. In some cases, for the undistributed users, the distribution destination users are selected within the empty range of the transmission queue. Thereby, while improving the utilization efficiency of a resource, the service improvement with respect to a user can be aimed at.

  In addition, according to the present embodiment, after the priority of the transmission queue is preferentially distributed to the priority content indicated by the priority content information, the remaining transmission queue is allocated to the popularity for the non-priority content. Distribute accordingly. Thereby, articles with high urgency such as earthquake early warning can be distributed preferentially.

  In addition, according to the present embodiment, the priority user indicated by the priority user information is preferentially selected as the distribution destination, and then the distribution destination users corresponding to the remaining number of distributions are selected. As a result, it is possible to preferentially distribute to users who subscribe to a paid distribution service.

In addition, the following effects can be obtained by this embodiment.
Some blog sites also have a time zone in which one or more posts are generated per second, and the number of newly arrived content may explode during such a time zone. In addition, there is an increasing need for browsing and utilizing such as extracting reputation information from articles on blog sites. Under such a blog site situation, by applying the content distribution control process according to the present embodiment, the load on the distribution server is properly controlled and appropriate new content is distributed to the user. It is possible to obtain an excellent effect of being able to.

  The distribution server 1 according to the present embodiment may be realized by dedicated hardware, or may be configured by a computer system such as a personal computer, and functions of each unit of the distribution server 1 shown in FIG. The function may be realized by executing a program for realizing the above.

Further, it is assumed that an input device, a display device, and the like (none of which are shown) are connected to the distribution server 1 as peripheral devices. Here, the input device refers to an input device such as a keyboard and a mouse. The display device refers to a CRT (Cathode Ray Tube), a liquid crystal display device or the like.
The peripheral device may be connected directly to the distribution server 1 or may be connected via a communication line.

Further, by recording a program for realizing the function of the distribution control unit 17 shown in FIG. 6 on a computer-readable recording medium, and causing the computer system to read and execute the program recorded on the recording medium, Content distribution control processing may be performed. Here, the “computer system” may include an OS and hardware such as peripheral devices.
“Computer-readable recording medium” refers to a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a DVD (Digital Versatile Disk), and a built-in computer system. A storage device such as a hard disk.

Further, the “computer-readable recording medium” means a volatile memory (for example, DRAM (Dynamic DRAM) in a computer system that becomes a server or a client when a program is transmitted through a network such as the Internet or a communication line such as a telephone line. Random Access Memory)), etc., which hold programs for a certain period of time.
The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

As mentioned above, although embodiment of this invention was explained in full detail with reference to drawings, the specific structure is not restricted to this embodiment, The design change etc. of the range which does not deviate from the summary of this invention are included.
For example, the maximum capacity of the transmission queue may be dynamically changed. An example of a calculation formula for the maximum capacity Qmax of the transmission queue is shown.
Qmax = T × Bs ÷ C
However, T is a period of content distribution control processing, Bs is a communication band that can be used by the distribution server 1 in the communication network 5 used at the time of content distribution, and C is an average value of the data amount of the content.

  Note that the functions of the respective units of the distribution server 1 shown in FIG. 2 may be implemented together in one device, or may be configured to be distributed and implemented in a plurality of devices.

It is a schematic structure figure showing the whole contents distribution system composition concerning one embodiment of the present invention. It is a block diagram which shows the structure of the delivery server 1 which concerns on the same embodiment. It is a structural example of the newly arrived content storage part 12 shown in FIG. It is the flowchart which showed the procedure of the feature vector calculation process which concerns on one Embodiment of this invention. It is a structural example of the participation information storage part 16 shown in FIG. It is a block diagram which shows the structure of the delivery control part 17 shown in FIG. It is the flowchart which showed the procedure of the content delivery control process which concerns on one Embodiment of this invention. It is the flowchart which showed the procedure of the content delivery control process which concerns on the same embodiment. It is the flowchart which showed the procedure of the content delivery control process which concerns on the same embodiment. It is the flowchart which showed the procedure of the content delivery control process which concerns on the same embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Distribution server, 2 ... Web (Web) site, 3 ... Internet, 4 ... User terminal, 5 ... Communication network, 11 ... New arrival content receiving part, 12 ... New arrival content storage part, 13 ... Feature vector calculation part, 14 ... Base category database (base category DB), 15 ... participation information receiving unit, 16 ... participation information accumulating unit, 17 ... distribution control unit (content distribution control device), 18 ... distribution unit, 21 ... content maximum distribution number calculating unit, 22 ... Content distribution number calculation unit, 23 ... First content distribution destination selection unit, 24 ... Second content distribution destination selection unit, 25 ... Preferential content information storage unit, 26 ... Preferential user information storage unit, 27 ... Non-distribution user information storage Part

Claims (7)

Content distribution that distributes the surplus of resources available for content distribution to new distribution target content according to the popularity of the content, and calculates the number of distributions for each new distribution target content based on the surplus resource distribution result Number calculation means;
First content distribution destination selection means for selecting a number of distribution destination users according to the number of distribution of the distribution target content based on the similarity between the user preference and the distribution target content grasped from the user's content browsing history; , equipped with a,
The content distribution number calculating means calculates a viewer rate for each of a plurality of similar contents distributed in the past similar to the new distribution target content from a user browsing history of the similar contents, and the calculated plurality of browsing A trend of increase / decrease between the plurality of viewer rates corresponding to the acquisition timing of the plurality of similar contents using a user rate, and the degree of popularity of the new distribution target content according to the calculated trend of increase / decrease Raise and lower,
A content distribution control apparatus characterized by the above. The user preference is calculated as a vector that combines the feature vectors of the content viewed by the user,
The content distribution control apparatus according to claim 1, wherein the first content distribution destination selection unit calculates a similarity between the vector of the calculation result and the feature vector of the new distribution target content. The first content delivery destination selection means, according to claim 1 or, characterized in that to reflect the participation rate of the user calculated from the content distribution history and the user content viewing history of the user in selecting the destination user 2. The content distribution control device according to 2. Undelivered user information storage means for storing information on undelivered users that are lost from selection of delivery destination users by the first content delivery destination selection means;
Second content distribution in which there is no new distribution target content and there is a surplus of resources that can be used for content distribution, and a distribution destination user is selected within the range of the surplus resource for undistributed users A destination selection means,
The content delivery control apparatus according to any one of claims 1 to 3 , further comprising: The content distribution number calculation means preferentially distributes the surplus of resources available for content distribution to the priority content indicated by the priority content information, and then distributes the remaining surplus resources to the non-priority content. content distribution control device according to distribute claim 1, wherein any one of claims 4 depending on the time. The first content distribution destination selection means selects a priority user indicated by priority user information as a distribution destination preferentially, and then selects distribution destination users for the remaining number of distributions. The content delivery control apparatus according to any one of claims 1 to 5 . Content distribution that distributes the surplus of resources available for content distribution to new distribution target content according to the popularity of the content, and calculates the number of distributions for each new distribution target content based on the surplus resource distribution result Number calculation function,
A first content distribution destination selection function for selecting the number of distribution destination users corresponding to the number of distribution of the distribution target content based on the similarity between the user preference and the distribution target content ascertained from the user's content browsing history; , A computer program for causing a computer to realize
The content distribution number calculation function calculates a viewer rate for each of a plurality of similar contents distributed in the past similar to the new distribution target content from a user browsing history of the similar contents, and the calculated plurality of browsing A trend of increase / decrease between the plurality of viewer rates corresponding to the acquisition timing of the plurality of similar contents using a user rate, and the degree of popularity of the new distribution target content according to the calculated trend of increase / decrease Raise and lower,
A computer program characterized by the above.

Images