JP2016173623A - Content providing device, content providing method, and content providing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the comprehensiveness of acquisition of logs related to an identical user, and provide a content suited to user's preference without mental burdens such as the leakage of personal information.SOLUTION: The following processing is performed on a user terminal (smart phone 1): collecting log information based on a user operation; generating a preference model indicating the degree of user's preference, by using the log information; and controlling the method for outputting a distribution content which is distributed from a server, by using the preference model.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a technology for providing content.

  There is a need for technology that provides services that suit each user. For this reason, currently, access logs to websites and user life logs are collected by a server, user preference characteristics are analyzed on the server side, and contents corresponding to the analysis results are transmitted to the user terminal. .

Takeuchi, 3 others, “Composite matrix decomposition under non-negative constraints”, IPSJ research report, mathematical modeling and problem solving research report, IPSJ, 2013-MPS-93 (3), 20135 16th of month

  However, since log information is accumulated in a plurality of servers with different operators, it is practically difficult to acquire all logs related to a predetermined user. Further, since the log itself corresponds to the user's personal information, the user's mental barrier is high against being stored and analyzed on the server side.

  The present invention has been made in view of the above circumstances, and improves the acquisition comprehensiveness of logs related to the same user, and provides content that matches the user's preference without causing a mental burden such as leakage of personal information. With the goal.

  In order to solve the above-described problem, the content providing apparatus according to claim 1 includes a collection unit that collects log information based on a user's operation at a user terminal, and a preference model that indicates the user's preference level using the log information. And a control means for controlling the output method of the distribution content distributed from the server using the preference model by the user terminal.

  According to the present invention, log information based on user operations is collected at the user terminal, so that it is possible to improve the acquisition comprehensiveness of logs related to the same user. In addition, since the process of collecting log information, the process of generating a user's preference model using the log information, and the process of controlling the delivery content output method using the preference model are performed on the user terminal, personal information It is possible to provide content that suits the user's preference without causing a mental burden such as leakage.

  The content providing device according to claim 2 further includes an adding unit that refers to the preference model and gives a score to the distribution content according to the content of the distribution content in the content providing device according to claim 1. Then, the gist of the invention is that the control means controls the output method of the distribution content according to the score value.

  A content providing apparatus according to a third aspect is the content providing apparatus according to the first or second aspect, further comprising transmission means for transmitting the preference model to another user terminal of the user.

  The content providing apparatus according to claim 4 is the content providing apparatus according to any one of claims 1 to 3, wherein an arbitrary preference model designated by the user among the plurality of preference models is transmitted to a backup server. It is a summary to further have means.

  The content providing device according to claim 5 is the content providing device according to any one of claims 1 to 4, wherein only the preference model shared with other user terminals of the user among the plurality of preference models. The gist of the present invention is to further include transmission means for transmitting the message to the backup server.

  The content providing method according to claim 6 is a content providing method performed by a content providing apparatus, a collecting step of collecting log information based on a user operation at a user terminal, and the user preference level using the log information. The present invention includes a generation step of generating a preference model to be shown at the user terminal, and a control step of controlling the output method of distribution content distributed from a server using the preference model at the user terminal.

  The content providing method according to claim 7 further includes an assigning step in which, in the content providing method according to claim 6, referring to the preference model, a score is given to the distribution content according to the content of the distribution content. Then, the gist of the control step is to control the delivery content output method according to the score value.

  A content providing program according to an eighth aspect of the present invention is to cause a computer to function as the content providing apparatus according to any one of the first to fifth aspects.

  ADVANTAGE OF THE INVENTION According to this invention, the acquisition comprehensiveness of the log regarding the same user can be improved, and the content suitable for a user's preference can be provided, without accompanying mental burden, such as leakage of personal information.

It is a reference figure for demonstrating the outline | summary of this invention. It is a figure which shows the functional block structure of a smart phone. It is a figure which shows the collection processing flow of log collection and a preference model. It is a figure which shows the recommendation process flow of a content. It is a figure which shows the functional block structure of a smart phone and a car navigation. It is a figure which shows the exchange and synchronous process flow of a preference model. It is a figure which shows the functional block structure of a personalization application and a backup server. It is a figure which shows the backup processing flow of a preference model. It is a figure which shows the restore processing flow of a preference model.

  First, the present invention will be outlined.

  In order to solve the above-described problem, the present invention collects a log based on a user operation on the user terminal side (smart phone 1) instead of the server side as shown in FIG. learn. The log is not sent to the server as before. When the content is distributed from the content distribution server 4 of the service provider, the user terminal analyzes and outputs the content that matches the user's preference by using the preference model.

  In addition, the present invention directly exchanges and synchronizes user preference models with other user terminals. This assumes a case where the preference model generated by one user terminal (smart phone 1) is diverted to other user terminals (car navigation system 2, digital television 3) of the same user. Since no server is interposed, the possibility of personal information remaining on the server side can be eliminated.

  Furthermore, the present invention transmits only a part of preference models to be shared to the backup server 5 only when a user preference model is shared by a plurality of user terminals (smart phone 1, car navigation system 2, digital television 3). . This assumes that the preference model is backed up and restored. Since all preference models are not transmitted, the possibility that personal information remains on the server side can be eliminated as much as possible.

  In addition, when the said some preference model is transmitted to servers (for example, advertisement production server 6), such as an advertising agency, smart phone 1 advertises offer price of a preference model via a physical means or an electronic means. Receive and receive from business operators. The advertising business here is an example of a transmission destination / sales destination of a preference model, and can target any individual user or company in general who uses the user model.

  Next, an embodiment for realizing the above-described present invention will be described with reference to the drawings.

[Log collection, preference model generation, etc.]
First, a configuration and operation for collecting logs, learning / generating preference models, and recommending contents in a user terminal will be described. FIG. 2 is a diagram illustrating a functional block configuration of the smartphone 1. The smartphone 1 includes a browser 11, a first application 12, a second application 13, a personalized application 14, and a service providing application 15.

  The browser 11 has a function for HTTP access to the web server 7. For example, an HTTP request transmission unit 11 a that transmits an HTTP request to the web server 7 and an HTTP response reception unit 11 b that receives an HTTP response from the web server 7 are configured. This can be realized by using a browser function attached to the smartphone 1.

  The first application 12 and the second application 13 are arbitrary applications installed in the smartphone 1. For example, an application log transmission unit 12a that transmits an application log generated during the operation of the first application 12 and an application log transmission unit 13a that transmits an application log generated during the operation of the second application 13 are provided. Configured. This can be realized by using an application capable of sending logs. The number of applications may be zero, one, or two or more. In the case of 0, the application is absent and the application log is not transmitted, but the preference model in this case is generated by using the HTTP log based on the access from the browser 11.

  In the present embodiment, the browser 11, the first application 12, and the second application 13 will be described as an example. However, since it is only necessary to collect logs based on user operations on the smartphone 1, it is connected to the outside of the smartphone 1. You may use what was done. For example, a biosensor or the like. In this case, a life log such as blood pressure is collected, which is also a kind of application log.

  The personalization application 14 has a function of collecting a log based on a user operation, learning and generating a user preference model from the log, and assigning a score based on the preference model to distribution content that has received a distribution request. ing.

  For example, an HTTP request receiving unit 14 a that receives an HTTP request from the browser 11 for the web server 7, an HTTP request transmitting unit 14 b that transmits the received HTTP request to the web server 7, and an HTTP response from the web server 7. An HTTP response receiving unit 14c and an HTTP response transmitting unit 14d that transmits the received HTTP response to the browser 11 are configured.

  In addition, an HTTP log collection unit 14e that collects HTTP requests and HTTP responses transmitted and received between the browser 11 and the web server 7 as an HTTP log, and the first application 12 and / or the second application 13 are generated. An application log receiving unit 14f that receives a log of the application as an application log, a model generation unit 14g that generates a preference model indicating a user's preference level using the HTTP log and / or the application log, and stores the preference model And a model database 14h.

  Further, the distribution content is referred to according to the content of the distribution content by referring to the score assignment request receiving unit 14i that receives a request for assigning a score to the distribution content and the user preference model stored in the model database 14h. A score assignment unit 14j for assigning a score to the content, and a score assignment result transmission unit 14k for transmitting a score assignment result for the distribution content.

  The service providing application 15 has a function of receiving the distribution content distributed from the content distribution server 4 and controlling the distribution content output method based on the score assignment result by the personalization application 14 and providing it to the user. For example, a content receiving unit 15a that receives distribution content, a score assignment requesting unit 15b that requests addition of a score to the received distribution content, and a score assignment result receiving unit 15c that receives a score addition result for the distribution content And a service providing unit 15d that adjusts a delivery content output method based on the received score assignment result and outputs the adjusted delivery content to a screen or a speaker.

  The personalization application 14 and the service providing application 15 can be realized by programming so as to realize the above-described functions using the CPU and the memory of the smartphone 1.

  In addition, although this Embodiment demonstrates as an example the case where each function part mentioned above is mounted in the smart phone 1, you may use the other user terminal provided with a communication function. For example, a mobile terminal, a tablet terminal, a personal computer, a game machine, a car navigation system, or a digital television may be used.

  Next, the operation of the smartphone 1 will be described. First, log collection and preference model generation operations will be described with reference to FIG.

  First, when the user operates the browser 11, in step S101, the personalization application 14 collects HTTP requests and / or HTTP responses transmitted / received between the browser 11 and the web server 7 as HTTP logs. Further, when the user uses the first application 12 and / or the second application 12, the application log transmitted from the corresponding application is received as the application log.

  Next, in step S102, the personalization application 14 mechanically learns the user's preference model using the collected and / or received logs (HTTP log, application log), and accumulates the generated preference model in the model database 14h. To do.

  Finally, in step S103, the personalization application 14 deletes the log collected or received in step S101.

  Thereafter, the process returns to step S101, and steps S102 to S103 are repeatedly executed every time a log is collected. Thereby, one or more preference models for the user who uses the smartphone 1 are learned and generated statistically and mechanically. Hereinafter, an example of generating a preference model will be described.

  For example, a searched word, a search frequency, an access destination store, a handling product, an access time, and the like are extracted from an HTTP log, and a user's preference for various arbitrary contents such as a product type is analyzed. Analysis may be performed in units of time or in units of days of the week. The degree of preference for each content may be analyzed and managed by the size and angle of the vector. Specifically, for example, when the user is searching for Chinese and Japanese food, a food preference model (Chinese: 40%, Japanese food: 60%) corresponding to the ratio of each search is generated. In addition, an arbitrary preference model such as a clothing preference model or a travel preference model may be generated based on collected and received logs. For example, when there is no HTTP log regarding a specific word, the preference tendency regarding the word may be set to 0 (default value).

  As a statistical / mechanical learning technique, for example, a non-negative multiple matrix factorization (NMMF) method using a framework for analyzing a plurality of data simultaneously described in Non-Patent Document 1 is used. In addition, a non-negative multiple tensor factorization (NMTF) method and a self-organizing incremental neural network (SONN) method may be used. Any technique that can model user preferences is acceptable.

  Next, the content recommendation operation will be described with reference to FIG. Note that the service provider wants to provide content that matches the user's preferences, but for that purpose, the service provider does not sort the content by itself. Therefore, it is assumed that various arbitrary contents are distributed regardless of the type and nature.

  First, in step S201, when the service providing application 15 receives a plurality of contents from the content distribution server 4 of the service provider, the service providing application 15 transmits a request for assigning scores to the received plurality of contents to the personalization application 14. .

  Next, in step S202, the personalization application 14 reads the user's preference model from the model database 14h, and gives a score to each distributed content. For example, when Chinese and Japanese content related to food is distributed, 4 points are given to the Chinese content and 6 points are given to the Japanese content, referring to the food preference model described above.

  Next, in step S <b> 203, the personalization application 14 transmits a score assignment result for each content to the service providing application 15.

  Finally, in step S204, the service providing application 15 adjusts the output method based on the assigned score, displays the content on the screen, and provides it to the user. For example, in the case of the above-described example, the content of Japanese food with high user preference (high score) is displayed first, and the content of Chinese with low preference (low score) is displayed next. Only high-scoring Japanese food content may be displayed, or voice conversion may be performed for voice output. Further, according to the score value, the character color, the character size, the blinking state, and the like may be changed, and the voice amount may be changed. Other arbitrary adjustment methods are conceivable.

[Exchange and synchronization of preference models]
Next, the configuration and operation for exchanging and synchronizing preference models between user terminals will be described. FIG. 5 is a diagram showing functional block configurations of the smartphone 1 and the car navigation system 2. In the personalized application 14, the smartphone 1 transmits an authentication processing unit 141 for authentication with another user terminal and a model transmission for transmitting the user preference model stored in the model database 14h to the other user terminal. 14m, a model receiving unit 14n that receives a preference model transmitted from another user terminal, a model update unit 14o that updates the preference model in the model database 14h with the received preference model, and another user terminal And a synchronization unit 14p that manages an opportunity for synchronizing the preference model. Since the car navigation 2 is also a kind of user terminal, the personalized application 14 of the car navigation 2 also has the same function unit.

  Next, a preference model exchange / synchronization operation will be described with reference to FIG. Here, the case where the preference model stored in the smartphone 1 is exchanged and synchronized with the car navigation 2 will be described as an example.

  First, in step S301, the personalized application 14 of the smartphone 1 executes an authentication process with the car navigation 2 using the user ID and the password.

  Next, when the authentication process is successful, in step S302, the personalized application 14 of the smartphone 1 reads the user's preference model from the model database 14h and transmits it to the car navigation system 2.

  Next, in step S303, the personalized application 14 of the car navigation system 2 receives the preference model transmitted from the smartphone 1, and stores it in the model database 14h at the timing of synchronization. At this time, the existing preference model may be replaced with the received preference model, or may be simply added. When adding, the existing preference model may be temporarily saved in other storage means. Alternatively, the preference database may be stored in the model database 14h when the preference model is received without waiting for the synchronization timing.

  The same process is performed when a preference model is transmitted from the car navigation 2 to the smartphone 1. Further, it may be performed in one direction or in both directions. Furthermore, when the user of the smartphone 1 and the car navigation 2 is the same, the authentication process in step S301 may be omitted. Moreover, you may use authentication means other than a user ID and a password. Furthermore, the communication path between the smartphone 1 and the car navigation system 2 may be encrypted, or the preference model itself may be encrypted. Furthermore, although personal information may remain on the server side, it may be transmitted via the server. In this case, it is possible to eliminate the necessity of starting both personalized applications 14 simultaneously.

[About backup / restoration of preference models]
Finally, the configuration and operation for performing backup / restoration of the preference model will be described. FIG. 7 is a diagram showing functional block configurations of the personalization application 14 and the backup server 5. FIG. 7A shows the functional configuration during backup, and FIG. 7B shows the functional configuration during restoration.

  Since the preference model is transmitted from the smartphone 1 in the same manner as the transmission process at the time of the exchange / synchronization process described above, the personalized application 14 has the same configuration as that shown in FIG. That is, the authentication processing unit 141 for authenticating with the backup server 5, the model transmission unit 14m for transmitting the user preference model to the backup server 5 for backup, and the backup server 5 for restoration. The model receiving unit 14n that receives the preference model and the model update unit 14o that updates the preference model of the model database 14h with the received preference model.

  The backup server 5 has a similar configuration. Specifically, the authentication processing unit 5a for authenticating with the smartphone 1, the model receiving unit 5b for receiving the preference model transmitted from the smartphone 1 for backup, and the received preference model as the model database 5d. A model updating unit 5c that stores the received preference model, a model database 5d that stores the received preference model, and a model transmission unit 5e that transmits the preference model that has been backed up for restoration to the smartphone 1. .

  Next, the preference model backup operation will be described with reference to FIG.

  First, in step S401, the personalization application 14 of the smartphone 1 executes an authentication process with the backup server 5 using a user ID and a password. Other authentication means may be used, and the communication path with the backup server 5 may be encrypted.

  Next, when the authentication process is successful, in step S402, the personalized application 14 of the smartphone 1 reads the user's preference model from the model database 14h, and communicates with other user terminals (car navigation system 2, digital television 3). Only the preference model to be shared specified by the user is transmitted to the backup server 5. At this time, all the generated preference models may be transmitted, or an arbitrary preference model may be selected and transmitted.

  Finally, in step S403, the backup server 5 receives the preference model transmitted from the smartphone 1, and accumulates (backups) it in the model database 5d.

  Subsequently, the restore operation of the preference model will be described with reference to FIG.

  First, in step S501, the personalized application 14 of the smartphone 1 executes an authentication process with the backup server 5 using a user ID and a password. Similar to the backup authentication, other authentication means may be used, and the communication path to the backup server 5 may be encrypted.

  Next, when the authentication process is successful, the backup server 5 reads the user preference model from the model database 5d and transmits it to the smartphone 1 in step S502.

  Finally, in step S503, the personalized application 14 of the smartphone 1 receives the preference model transmitted from the backup server 5 and accumulates (restores) it in the model database 14h.

  As mentioned above, according to this Embodiment, since the log based on a user's operation is collected by a user terminal (smart phone 1), the acquisition comprehensiveness of the log regarding the same user can be improved. There is no need to prepare a server for collecting the logs of a plurality of users in order to grasp the preferences of each user, and there is no need to perform a process of acquiring logs of various servers accessed by the users.

  Further, according to the present embodiment, a process of collecting log information, a process of generating a user's preference model using the log information, and a process of controlling a delivery content output method using the preference model. Since it is performed by the user terminal (smart phone 1), it is possible to provide content that meets the user's preference without causing a mental burden such as leakage of personal information. Moreover, since all logs of the same user can be collected, it is possible to provide content that is more suitable for the user's preference than before.

DESCRIPTION OF SYMBOLS 1 ... Smartphone 11 ... Browser 11a ... HTTP request transmission part 11b ... HTTP response reception part 12 ... 1st application 12a ... Application log transmission part 13 ... 2nd application 13a ... Application log transmission part 14 ... Personalized application (content provision apparatus) )
14a ... HTTP request reception unit 14b ... HTTP request transmission unit 14c ... HTTP response reception unit 14d ... HTTP response transmission unit 14e ... HTTP log collection unit 14f ... Application log reception unit 14g ... Model generation unit 14h ... Model database 14i ... Score assignment request Accepting unit 14j ... score assigning unit 14k ... score assigning result sending unit 14l ... authentication processing unit 14m ... model sending unit 14n ... model receiving unit 14o ... model updating unit 14p ... synchronizing unit 15 ... service providing application (content providing device)
DESCRIPTION OF SYMBOLS 15a ... Content reception part 15b ... Score provision request | requirement part 15c ... Score provision result reception part 15d ... Service provision part 2 ... Car navigation system 3 ... Digital television 4 ... Content delivery server 5 ... Backup server 5a ... Authentication process part 5b ... Model reception part 5c ... model update unit 5d ... model database 5e ... model transmission unit 6 ... advertisement generation server 7 ... web server S101 to S103, S201 to S204, S301 to S303, S401 to S403, S501 to S503 ... step

In order to solve the above-described problem, the content providing apparatus according to claim 1 includes a collection unit that collects log information based on a user's operation at a user terminal, and a preference model that indicates the user's preference level using the log information. Generating means for generating at the user terminal, and receiving any distribution content distributed from the server without transmitting information about the user's preference level to the server. used and summarized in that with the control Gosuru control means an output method of the distribution content.

The content providing apparatus according to claim 4, wherein in the content providing apparatus according to claim 1 or 2 , transmission means for transmitting an arbitrary preference model designated by the user to the backup server among the plurality of preference models. Furthermore, it is summarized as having.

The content providing device according to claim 5 is a backup server according to claim 1 or 2 , wherein only the preference model shared among the user's other user terminals among the plurality of preference models is a backup server. The gist of the present invention is to further include transmission means for transmitting to the network.
The content providing apparatus according to claim 6 is the content providing apparatus according to any one of claims 1 to 5, wherein the generation means is a predetermined statistical learning method based on information extracted from the log information. Alternatively, the gist model is to generate the preference model by a mechanical learning method.
A content providing device according to a seventh aspect is the content providing device according to any one of the first to sixth aspects, wherein the collection unit deletes the log information.

The content providing method according to claim 8 , in the content providing method performed by the content providing device, a collection step of collecting log information based on a user operation at a user terminal, and the user preference level using the log information. In the user terminal, a generation step of generating a preference model to be shown at the user terminal, and receiving any distribution content distributed from the server without transmitting information on the user preference level to the server, and summarized in that with the control Gosuru control step the output method of the distributed content using the preference model.

The content providing method according to claim 9 further includes an assigning step of referring to the preference model and assigning a score to the distribution content according to the content of the distribution content in the content providing method according to claim 8. Then, the gist of the control step is to control the delivery content output method according to the score value.
The content providing method according to claim 10 is the content providing method according to claim 8 or 9, wherein, in the generation step, based on information extracted from the log information, a predetermined statistical learning method or mechanical The gist is to generate the preference model by a learning method.
A content providing method according to an eleventh aspect is the content providing method according to any one of the eighth to tenth aspects, wherein the log information is deleted in the collecting step.

A content providing program according to a twelfth aspect causes a computer to function as the content providing apparatus according to any one of the first to seventh aspects.

Claims (8)

A collecting means for collecting log information based on user operations at the user terminal;
Generating means for generating, at the user terminal, a preference model indicating the degree of preference of the user using the log information;
Control means for controlling the output method of the delivery content delivered from the server using the preference model at the user terminal;
A content providing apparatus comprising: Further comprising an assigning unit that refers to the preference model and gives a score to the distribution content according to the content of the distribution content,
The control means includes
The content providing apparatus according to claim 1, wherein an output method of the distribution content is controlled according to the score value. Transmitting means for transmitting the preference model to another user terminal of the user;
The content providing apparatus according to claim 1, further comprising: Transmitting means for transmitting an arbitrary preference model designated by the user among the plurality of preference models to a backup server;
The content providing apparatus according to claim 1, further comprising: A transmitting means for transmitting only a preference model to be shared with other user terminals of the user among the plurality of preference models to the backup server;
The content providing apparatus according to claim 1, further comprising: In the content providing method performed by the content providing device,
A collection step of collecting log information based on user operations at the user terminal;
Generating a preference model indicating the user's preference level using the log information at the user terminal;
A control step of controlling, at the user terminal, an output method of distribution content distributed from a server using the preference model;
A content providing method characterized by comprising: A step of referring to the preference model and adding a score to the distribution content according to the content of the distribution content;
In the control step,
The content providing method according to claim 6, wherein an output method of the distribution content is controlled according to the score value.   A content providing program for causing a computer to function as the content providing apparatus according to claim 1.

Images