JP2013164792A - Information processing device, information processing method, and program - Google Patents

Abstract

An information processing apparatus capable of providing information suitable for a user's preference to a user without a usage history is provided.
A storage unit that stores preference information for each existing user, and human relation information that is information indicating a person-to-person connection for a newly registered user is acquired from another information processing apparatus via a network. Control that identifies an existing user associated with a newly registered user by referring to the relationship information, reads the specified preference information of the existing user from the storage unit, and estimates the preference information of the newly registered user based on the read preference information Part.
[Selection] Figure 1

Description

  The present invention relates to an information processing apparatus that provides information to a user via a network, an information processing method, and a program for causing the information processing apparatus to execute the method.

  Information providing sites such as social network service (SNS) sites, search websites, and news providing sites have come to be used by many people via the Internet.

  In these information providing sites, information recommendation using user preferences is performed. If the user himself / herself operates an information terminal to directly input information related to his / her preference in order to recommend information according to the user's preference, the user is troublesome. In consideration of user convenience, collaborative filtering that extracts the degree of association from the usage history and content information for each user and recommends information suitable for the user is becoming mainstream (see Non-Patent Document 1).

  In Amazon (registered trademark) of EC (Electric Commerce) site, when recommending a product to a user, a user with a similar preference is estimated based on the user's product purchase history and recorded in the estimated product purchase history of the user I recommend products. In addition, in iTunes (registered trademark), which is a media player of Apple (registered trademark), when recommending a song to a user, a user who has a close preference based on information on the song possessed by the user, such as the user's song purchase history And the music purchased by the extracted user is recommended.

Noriaki Kawamae, “Let's Make a Recommendation System with Hadoop”, [online], October 18, 2011, Technical Review, [Search February 7, 2012], Internet <URL: http: // gihyo. jp / dev / serial / 01 / recommend_hadoop / 0001>

  In the information recommendation method using the use history such as the purchase history of the user, in the case of a user newly registered as a member in a certain EC site, the EC site has no use history and the information recommendation is immediately made to the user. I can't. In particular, when a user registers on the EC site for the purpose of receiving information recommendation, the information recommendation cannot be received immediately even after registration.

  In order to solve this problem, if preference information that is information related to the preference of the same user is accumulated in another information providing site, a method of using the preference information may be considered. However, the preference information is complicated, and if it is constructed with different contents and formats for each information providing site, the preference information of other information providing sites cannot be used as it is. In addition, for users, preference information and usage history may correspond to fresh information representing one's own personality, living environment, etc., and there is high resistance to disclosing such information to other information providing sites, In some cases, it is necessary to obtain a user's permission when receiving provision of preference information.

  The present invention has been made in order to solve the problems of the above-described technology, and an information processing apparatus and an information processing apparatus that can provide information suitable for a user's preference to a user who has no usage history It is an object to provide a method and a program for causing an information processing apparatus to execute the method.

In order to achieve the above object, an information processing apparatus of the present invention provides:
A storage unit for storing preference information for each existing user who is an already registered user;
For a newly registered user who is a newly registered user, human relation information that is information indicating a person-to-person connection is acquired from another information processing apparatus via the network, and the new relation user is referred to and the new relation user is referred to. A control unit that identifies an existing user associated with a registered user, reads the preference information of the existing user from the storage unit, and estimates the preference information of the newly registered user based on the read preference information;
It is the structure which has.

An information processing method of the present invention is an information processing method executed by an information processing apparatus having a storage unit and a control unit,
The control unit prestores preference information for each existing user who is a registered user in the storage unit,
When the user is newly registered, the control unit transmits, to the other information processing apparatus, the human relationship information, which is information indicating a person-to-person connection, for the newly registered user who is a newly registered user via the network. Get from
The control unit identifies an existing user associated with the newly registered user with reference to the human relationship information,
The control unit reads the specified preference information of the existing user from the storage unit, and estimates the preference information of the newly registered user based on the read preference information.

Furthermore, the program of the present invention is provided in an information processing apparatus.
A procedure for storing the preference information for each existing user who is already registered in the storage unit in advance;
A procedure for acquiring human relationship information, which is information indicating a person-to-person relationship, from another information processing apparatus via a network for a newly registered user who is a newly registered user when a user is newly registered; ,
Identifying an existing user associated with the newly registered user with reference to the human relationship information;
The specified preference information of the existing user is read out from the storage unit, and the procedure of estimating the preference information of the newly registered user based on the read preference information is executed.

  According to the present invention, even a user who has almost no usage history can provide information that meets the user's preference, and the convenience of the user can be improved.

It is a block diagram which shows the example of 1 structure of the communication system containing the information processing apparatus of this embodiment. 1 is a block diagram illustrating a configuration example of a communication system according to a first embodiment. FIG. 3 is a block diagram illustrating a configuration example of an SNS server illustrated in FIG. 2. It is a table | surface which shows an example of user information. 6 is a table illustrating an example of user relationship information in the first embodiment. It is a figure which shows an example of the graph which visualized the user relationship information shown in FIG. FIG. 3 is a block diagram illustrating a configuration example of an EC server illustrated in FIG. 2. It is a table | surface which shows an example of a service utilization log | history. It is a table | surface which shows an example in case preference information is calculated with a vector. It is an example of the user relationship information received from an SNS server. It is a block diagram which shows one structural example of the user terminal shown in FIG. FIG. 6 is a schematic diagram for explaining an operation of the EC server according to the first embodiment. FIG. 3 is a sequence diagram illustrating an operation procedure of the communication system according to the first embodiment. It is a flowchart which shows the operation | movement procedure of EC server which switches the information to be used in the case of the information recommendation with respect to a user. It is a block diagram which shows one structural example of EC server of Example 2. FIG. FIG. 10 is a schematic diagram for explaining an operation of the EC server according to the second embodiment. It is a figure which shows an example of the user relationship information received from the SNS server. It is a figure which shows an example of the user relationship information received from another SNS server. It is a figure which shows the user relationship information which combined FIG. 17A and FIG. 17B. FIG. 10 is a block diagram illustrating a configuration example of a communication system according to a fourth embodiment. It is a schematic diagram for demonstrating the structure of the SNS server of Example 4, and EC server.

  The information processing apparatus of this embodiment will be described. FIG. 1 is a block diagram illustrating a configuration example of a communication system including the information processing apparatus according to the present embodiment.

  As illustrated in FIG. 1, the communication system includes an information processing device 2 and an information processing device 4 connected to a network 5. User terminal 3 is connected to network 5. The information processing apparatus 4 is a server that provides SNS to the user. The information processing apparatus 2 is a server that performs information processing for selling products to the user. The information processing apparatus 4 includes an SNS function for building a human relationship on the network, and holds the human relationship information accumulated by the SNS function. The information processing apparatus 2 holds in advance preference information for each registered user.

  When a user is newly registered via the user terminal 30, the information processing apparatus 2 is a person who is information indicating a person-to-person connection for a newly registered user (hereinafter referred to as a newly registered user). The relationship information is acquired from the information processing device 4 via the network 5, the other user associated with the newly registered user is identified by the human relationship information, and the preference of the newly registered user is determined based on the identified preference information of the other user. Estimate information.

  In the present embodiment, human relationship information of a newly registered user is used for estimating preference information of the newly registered user. Since it is predicted that the users having relationships with human relations are close to each other, it is possible to estimate the preferences of the newly registered users by using the preference information of other users having the relationship of the newly registered users. . As a result, even a newly registered user who has no usage history of the information processing apparatus 2 can be provided with recommended information close to his / her preference.

In addition, human relationship information is simpler and less blurred than preference information, has a well-established format, and is often used to search for known friends on newly registered sites. There is little resistance to disclosure on the site.
Examples of the information processing apparatus according to the present embodiment will be described below.

  A configuration of a communication system including the information processing apparatus according to the present embodiment will be described.

  FIG. 2 is a block diagram illustrating a configuration example of the communication system according to the present embodiment. The EC server 20 shown in FIG. 2 corresponds to the information processing apparatus 2 shown in FIG.

  As shown in FIG. 2, the communication system includes an SNS server 10 and an EC server 20 that are connected to a network 50. The user terminal 30 is connected to the SNS server 10 or the EC server 20 via the network 50. The network 50 is, for example, the Internet. The SNS server 10 is a server that provides SNS to users. The EC server 20 is a server that performs information processing for selling products to users.

  Examples of Web sites established by the SNS server 10 in the network 50 include FACEBOOK (registered trademark) and mixi (registered trademark). Hereinafter, a website that the SNS server 10 establishes in the network 50 is referred to as an “SNS site”. Examples of Web sites that the EC server 20 establishes in the network 50 include Amazon (registered trademark) and Rakuten (registered trademark). Hereinafter, a Web site that the EC server 20 establishes in the network 50 is referred to as an “EC site”.

  Next, each configuration shown in FIG. 2 will be described in detail. In this embodiment, it is assumed that there are four users A to D who operate the user terminal 30 in order to explain the contents of information processing in an easy-to-understand manner. In addition, it is assumed that the users A to D are registered in advance in the SNS server 10. However, it is assumed that the users B to D are registered in advance in the EC server 20, but the user A has not been registered yet.

  FIG. 3 is a block diagram illustrating a configuration example of the SNS server.

  As illustrated in FIG. 3, the SNS server 10 includes a storage unit 13 and a control unit 11. The storage unit 13 stores user information 131 that is user-specific information and user relationship information 133 that is information indicating a connection between users. User information 131 and user relationship information 133 shown in FIG. 3 schematically represent storage areas in which the respective information is stored.

  The user information 131 is personal information including identification information (Identifier) that is different for each user, such as a user's name or nickname, an account, and a password. The user identification information is referred to as a user ID. User information 131 of users A to D is registered in the storage unit 13. The user relationship information 133 includes human relationship information that is information indicating a connection between people. There is what is called social graph and interest graph as human relation information. In this embodiment, it is assumed that the user relationship information 133 is a social graph.

  FIG. 4 is a table showing an example of user information. As shown in FIG. 4, personal information such as an e-mail address, sex, and age is described corresponding to the user ID. SNS2 and SNS3 mean SNS sites established by an SNS server (not shown) connected to the network 50. In the user information shown in FIG. 4, the IDs used in these SNS sites are also registered. Although shown, these IDs may not be registered.

  FIG. 5 is a table showing an example of the user relationship information, and shows the mutual connection between the users A to C. FIG. 6 is a diagram showing an example of a graph visualizing the user relation information shown in FIG.

  FIG. 6 shows that user A is connected to user B by a path, and user A has a direct connection with user B. Further, FIG. 6 shows that the user B is connected to the user C by the path, and the user B is directly connected to the user C. From the human relationship diagram shown in FIG. 6, it can be seen that user A is not directly related to user C, but is related to user C via user B.

  In the storage unit 13, in addition to the above various information, web page data to be displayed on the user terminal 30 is stored.

  The control unit 11 includes an information providing unit 111 that provides user relationship information to the EC server 20. The control unit 11 includes a CPU (Central Processing Unit) (not shown) that executes processing according to a program, and a memory (not shown) that stores the program. The information providing unit 111 is virtually configured in the SNS server 10 by the CPU executing the process according to the program. The information providing unit 111 functions as a user relationship information providing API (Application Program Interface) described later.

  When the information providing unit 111 receives from the EC server 20 a user relationship information request signal including information on requesting user relationship information and user ID from the EC server 20, the information providing unit 111 stores user relationship information 133 regarding the user specified by the user ID from the storage unit 13. Read and return to the EC server 20.

  Next, the configuration of the EC server 20 will be described. FIG. 7 is a block diagram showing a configuration example of the EC server.

  As shown in FIG. 7, the EC server 20 includes a storage unit 23 and a control unit 21. The storage unit 23 includes user information 231 that is user-specific information, product master information 232, service usage history 234, preference information 235, and estimated preference information 236. User information 231, product master information 232, service usage history 234, preference information 235, and estimated preference information 236 shown in FIG. 7 schematically represent storage areas in which the respective information is stored.

  The user information 231 is personal information including identification information that differs for each user, such as the user's name, account, and password. User information 231 of users B to D is registered in the storage unit 23. In this embodiment, it is assumed that the table shown in FIG. 4 excluding the row corresponding to the user A is registered in the storage unit 23 as the user information 231. Compared with the user information 131, there is a difference that the user information 231 includes at least information on the real name and address of the user in order to reliably deliver the product to the user.

  The product master information 232 is information related to products that can be provided to the user. For example, if the product is a classical music CD, the product master information 232 includes a product name, product ID and price information of “Mozart Requiem”. When the merchandise is sportswear, the merchandise master information 232 includes information on the merchandise name, the merchandise ID, and the price of “Japan soccer representative replica jersey”.

  The service usage history 234 is information related to products purchased by the user, and data is accumulated for each user. Corresponding to each of the users B to D, information on purchased products is registered as a service usage history 234. For example, when the user B purchases a “Mozart Requiem CD”, the information of the “Mozart Requiem CD” is registered as the service usage history 234 corresponding to the user B, and the user C is “Soccer Japan Representative Replica Jersey”. When “” is purchased, the information of “Soccer Japan Representative Replica Jersey” corresponding to the user C is registered as the service use history 234.

  FIG. 8 is a table showing an example of service usage history. From the service usage history shown in FIG. 8, it can be seen that the user B purchased one product specified by the product ID 123 at 10:10 on November 1, 2011.

  The preference information 235 is information related to user preferences and is registered for each user. Preference information 235 is registered in the storage unit 23 corresponding to each of the users B to D. For example, the preference information 235 may be classified according to a genre such as music or sports, but may be represented in a state that cannot be understood even by a person.

  The case where the preference information 235 is mechanically calculated as a vector corresponds to the case where the preference information 235 is expressed in a state that humans cannot see and understand. FIG. 9 is a table showing an example in which the preference information is calculated as a vector, and each of the preference information of the users B to D is expressed by an n-dimensional vector (n is an integer of 1 or more). The preference information 235 is not limited to the case where the preference information is classified by genre or the case where the preference information is calculated as a vector, and the expression format is not limited.

  The storage unit 23 stores web page data to be displayed on the user terminal 30 in addition to the above-described various types of information.

  The control unit 21 includes a preference information extraction unit 213, a preference information estimation unit 215, and an information recommendation unit 217. The control unit 21 includes a CPU (not shown) that executes processing according to a program, and a memory (not shown) that stores the program. As the CPU executes processing according to the program, the preference information extraction unit 213, the preference information estimation unit 215, and the information recommendation unit 217 are virtually configured in the EC server 20.

  The preference information extraction unit 213 extracts the preference information 235 estimated from the service usage history for each user, and stores the extracted preference information 235 in the storage unit 23.

  When the user A is newly registered, the preference information estimation unit 215 transmits a user relationship information request signal including the user ID of the user A to the SNS server 10. Then, when the preference information estimation unit 215 receives the user relationship information of the user A from the SNS server 10, the preference information estimation unit 215 reads the preference information 235 of another user associated with the user A with reference to the user relationship information from the storage unit 23, Based on the read preference information 235, the preference information of the user A is estimated and stored in the storage unit 23. The estimated preference information corresponds to the estimated preference information 236 shown in the figure.

  FIG. 10 shows an example of user relationship information received from the SNS server. FIG. 10 shows that the user A is directly associated with each of the users B to D. In addition, the user relation information shown in FIG. 10 includes information on the mail addresses of the users B to D as information for specifying other users associated with the user A in the EC server 20. In this embodiment, a mail address is used as information for specifying the users B to D, but information other than the mail address may be used.

  The information recommendation unit 217 extracts a product from the product master information 232 based on the preference information 235 or the estimated preference information 236 for each user, and provides the extracted product information to the user as recommendation information.

  Next, the configuration of the user terminal 30 will be described.

  The user terminal 30 is a kind of information processing terminal, and is not limited to a desktop personal computer, and may be a mobile terminal including a smartphone and a mobile terminal such as a PDA (Personal Digital Assistant).

  FIG. 11 is a block diagram illustrating a configuration example of a user terminal. As illustrated in FIG. 11, the user terminal 30 includes a storage unit 33, a display unit 35, an operation unit 37, and a control unit 31. As represented by a smartphone, a touch panel having both functions of the display unit 35 and the operation unit 37 may be provided in the user terminal 30. The control unit 31 includes a CPU (not shown) that executes processing according to a program, and a memory (not shown) that stores the program.

  The storage unit 33 stores a browser. When an instruction to connect to the SNS server 10 or the EC server 20 is input by the user via the operation unit 37, the control unit 31 reads and starts the browser from the storage unit 33, and starts the web page specified by the user. When the data is acquired from the connection destination device, the web page is displayed on the display unit 35.

  As for the data communication method between devices, when the network 50 is the Internet, data can be transmitted and received between the devices via the network 50 according to TCP (Transmission Control Protocol) / IP (Internet Protocol). In the present embodiment, detailed description of the data communication method is omitted.

  Next, the operation of the EC server 20 will be described. Here, a case where user A registers as a new member of an EC site established by the EC server 20 will be described.

  FIG. 12 is a schematic diagram for explaining the operation of the EC server of this embodiment. FIG. 13 is a sequence diagram showing an operation procedure of the communication system of the present embodiment.

  FIG. 12 is a diagram for explaining the flow of information in an easy-to-understand manner from the viewpoint of a user newly registered in the EC site. The user relationship information provision API 15 located between the SNS site and the EC site corresponds to the information provision unit 111 described with reference to FIG.

  When the user A operates the operation unit 37 of the user terminal 30 and inputs an instruction to connect to the EC server 20, the control unit 31 acquires homepage data from the EC server 20 via the network 50, When activated, the home page is displayed on the display unit 35. Subsequently, when the user A operates the operation unit 37 and selects “new member registration” from the web page displayed on the display unit 35, the control unit 31 acquires a web page for new registration from the EC server 20. .

  In addition to the column for entering the user's personal information, select whether or not to "disclose your user relationship information on the SNS site to this EC site" on the new registration web page There are items to do this. By providing this item, it is possible to cope with users who are reluctant to disclose their user relationship information to the EC site. The method of obtaining this permission from the user is not limited to the method of confirming on the new registration web page.

  When the control unit 31 of the user terminal 30 displays a web page for new registration on the display unit 35, the user A operates the operation unit 37 to input personal information such as his / her name. Here, it is assumed that the user A wants to immediately receive recommendation information related to the product, and operates the operation unit 37 to permit “disclosure of his / her user relationship information on the SNS site to the EC site”. Select the direction. When the user A completes the input of the new registration web page, the control unit 31 of the user terminal 30 receives the new registration information including information indicating that the disclosure of the personal information of the user A and the user relation information is permitted. (Step 1001).

  When receiving the new registration information from the user terminal 30, the control unit 21 of the EC server 20 registers user information based on the personal information of the user A in the storage unit 23 (step 1002). Since user A is a newly registered user and there is no EC site usage history, the preference information extraction unit 213 cannot extract the preference information of user A. Therefore, when the preference information estimation unit 215 of the control unit 21 confirms the information indicating that the disclosure of the user relationship information is permitted with the new registration information, the user ID of the user A is requested to request the user relationship information about the user A. Is transmitted to the SNS server 10 (step 1003).

  When the SNS server 10 receives the user relationship information request signal from the EC server 20, the information providing unit 111 reads the user relationship information regarding the user A from the storage unit 13, and transmits the user relationship information to the EC server 20 (step 1004). ). Here, the user relationship information is returned from the SNS server 10 to the EC server 20 in the format shown in FIG.

  When the user relationship information as shown in FIG. 10 is received from the SNS server 10, the preference information estimation unit 215 of the EC server 20 receives the mail address included in the user relationship information, the user ID in the SNS server 10, and the storage unit 23. Are associated with the mail address and user ID registered in the user information 231. Thereby, the preference information estimation unit 215 specifies the users B to D registered in the user information 231 of the storage unit 23 corresponding to the users B to D registered in the SNS server 10, and specifies the specified users B to D. And user A are linked according to the user relationship information (step 1005). In the example illustrated in FIG. 10, the user A has a relationship that is directly associated with each of the users B to D.

  Subsequently, the preference information estimation unit 215 uses the user relation information of the user A and the preference information of each of the users B to D as input data, estimates the preference information of the user A using machine learning, and calculates the estimated preference information. The information is output to the information recommendation unit 217 (step 1006). Here, a specific example of machine learning in step 1006 will be described below.

  The preference information estimation unit 215 uses the user relation information and preference information of existing users (here, users B to D) of the EC site as teacher data, and learns preference information based on the user relation information in advance. The learning result here is information such as selection of an estimator, an estimation parameter, and an estimation algorithm for estimating the preference information of the user A learned based on the teacher data. These learning results are stored in the storage unit 23. The machine learning platform used is not limited to a real-time processing platform such as Jubatus, but may be a batch processing platform. An example of learning data is shown below.

Learning data example input data: user m user relation information, user preference information related to the user m correct answer data: user m preference information where m is a code meaning any existing user.

  A specific example in the case where the preference information of the user A is estimated from the preference information of the users B to D will be described. Here, in order to facilitate understanding, the preference information is information that can be understood by humans.

  When user B purchases “spike shoes” using the EC site, since the spike shoes are used in sports, “sports” and “shoes” are registered as user B's preference information. The When the user C has purchased the “Japan soccer representative replica jersey” using the EC site, “soccer” and “jersey” are registered as the preference information of the user C. When the user D has purchased “soccer stockings” using the EC site, “soccer” and “stockings” are registered as the preference information of the user D. In this case, since soccer is a kind of sport, the preference information estimation unit 215 refers to the preference information of each of the users B to D related to the user A, and the “sport” of the user B and the “ “Soccer”, which is common to “soccer” and “soccer” of user D, is estimated as the preference information of user A.

  Moreover, the user A's preference information may be estimated from the service usage history of the users B to D using the service usage history of the users B to D as the respective preference information. Explaining in the above example, the preference information estimation unit 215 includes “spike shoes” purchased by the user B, “soccer Japan representative replica jersey” purchased by the user C, and service usage histories of the users B to D, and From the information of “soccer stockings” purchased by the user D, “soccer” is estimated as the preference information of the user A.

  Based on the estimated preference information 236 of user A output in step 1006 shown in FIG. 13, the information recommendation unit 217 of the EC server 20 extracts products from the product master information, and uses the extracted product information as recommendation information. It transmits to the user terminal 30 (step 1007). When receiving the recommendation information from the EC server 20, the control unit 31 of the user terminal 30 displays the recommendation information on the display unit 35 (step 1008).

  In order to improve the accuracy of the estimation process in step 1006, the following process may be performed. Each time the preference information of these users is updated by the users B to D using the EC site, the preference information estimation unit 215 executes the estimation process of the preference information of the user A, and based on the result. User A's estimated preference information is updated. What is necessary is just to make the CPU (not shown) of the control part 21 run the program which can be machine-learned with respect to the sequential update of such teacher data.

  In the EC server 20, the preference information estimation unit 215 estimates the preference information of the user A with the preference information of the users B to D until the service usage history of the user A is sufficiently accumulated, and the result is the estimated preference information 236. The preference information extraction unit 213 stores the preference information 235 of the user A estimated from the service usage history of the user A in the storage unit 23. In this manner, the preference information extraction unit 213 and the preference information estimation unit 215 are estimated until the service usage history of the user A is sufficiently accumulated, but the preference information estimated by the preference information extraction unit 213 is used. The estimated preference information 236 by the preference information estimation unit 215 is predicted to be closer to the user A's preference than the user A's preference.

  Here, an example of a method for switching the information to be used from “user A's estimated preference information 236” to “preference information 235 based on user A's service usage history” when the information recommendation unit 217 obtains user A's recommended information. Will be explained.

  FIG. 14 is a flowchart showing an operation procedure of the EC server for switching the information to be used from the estimated preference information to the preference information when recommending information to the user. Here, with respect to the accumulated number of service usage histories, a threshold value serving as a reference for timing of switching information used for extracting recommended information is determined in advance. Information on the threshold is registered in advance in a memory (not shown) in the control unit 21.

  As shown in FIG. 14, the control unit 21 determines whether or not the user A is connected to the user terminal 30 of the user A (step 1101). When connected to the user terminal 30 in a communicable manner, the control unit 21 refers to the service usage history for the user A of the user terminal 30 and determines whether or not the accumulated number of service usage histories is equal to or greater than a threshold (step 1102 ).

  If it is determined in step 1102 that the accumulated number of service usage histories is less than the threshold, the information recommendation unit 217 obtains and outputs recommendation information based on the estimated preference information of the user A (step 1103). On the other hand, if it is determined in step 1102 that the accumulated number of service usage histories is equal to or greater than the threshold, the information recommendation unit 217 obtains and outputs recommended information based on the preference information of the user A (step 1104).

  Thereafter, when the user A purchases a product (step 1105), the control unit 21 adds the information to the service usage history of the user A and updates the service usage history (step 1106). When the connection with the user terminal 30 is disconnected (step 1107), the control unit 21 ends the process.

  The method for switching the information used when the information recommendation unit 217 obtains the recommendation information of the user A from the estimated preference information 236 to the preference information 235 is not limited to the method illustrated in FIG. You may make it switch by a method.

  According to the information processing apparatus of the present embodiment, even if a user has almost no usage history on the EC site, other users related to the user based on the prediction that the users having a human relationship are close to each other. The preference information is estimated based on the user's preference information. For this reason, the user can be provided with recommended information close to his / her preference even if he / she has just registered on the EC site without providing information on his / her preference to the EC site himself / herself.

  The present embodiment is a case where the user relation information uniquely held by the EC server and the user relation information provided from the SNS server are used in combination. In the present embodiment, detailed description of the same configuration and operation as in the first embodiment will be omitted, and differences from the first embodiment will be described in detail.

  The configuration of the information processing apparatus according to this embodiment will be described. FIG. 15 is a block diagram showing a configuration example of the EC server of this embodiment. The EC server 22 shown in FIG. 15 corresponds to the information processing apparatus 2 shown in FIG.

  Music sale sites such as iTunes (registered trademark) are known as EC sites established by the EC server 22 of this embodiment. iTunes (registered trademark) provides a user with a site “Ping” having an SNS function. As shown in FIG. 15, the EC server 22 holds user relationship information 233 in the storage unit 23.

  The EC server 22 holds unique user relation information, but by receiving provision of other user relation information from the SNS server 10, more detailed user relation information can be used.

  When the EC site is a music sales site with an SNS function as described above, a user newly registered in the EC site searches for a person who has the same musical hobby as his own, The motivation for disclosing user-related information to the EC server 22 is further strengthened.

  Next, the operation of the EC server 22 of this embodiment will be described with reference to FIGS.

  FIG. 16 is a schematic diagram for explaining the operation of the EC server of this embodiment. 17A to 17C are diagrams illustrating an example of user relationship information. FIG. 17A is user relation information held by the SNS server 10, and FIG. 17B is user relation information held by the EC server 22.

  When the EC server 22 receives the user relation information shown in FIG. 17A from the SNS server 10 in step 1004 shown in FIG. 13, the preference information estimation unit 215 performs the user relation information shown in FIG. 17A and the user relation shown in FIG. 17B. The information is added to construct the user relationship information shown in FIG. 17C. Then, using the user relation information shown in FIG. 17C, the EC server 22 performs processing in the procedure after step 1005 shown in FIG. Note that the human relationship information obtained by adding a plurality of human relationship information as illustrated in FIG. 17C is referred to as “combined human relationship information”.

  In this embodiment, when the EC server 22 estimates the preference information of the user A, not only the users B to D to which the user A is directly related, but also the users E and users who are related to the users B to D, respectively. Information on F can also be used.

  Further, by using a plurality of user relationship information, the relationship between the user C, the user D, the user E, and the user F is deep from the number of paths shown in the user relationship information in FIG. It can be assumed that the relationship is deep. By using the information on the depth of the human relationship for the estimation process, it is possible to improve the estimation accuracy of the preference information.

  In the first embodiment, the user relation information acquisition method has been described in the case of the pull format in which the EC server requests user relation information regarding a specific user from the SNS server and is pulled from the SNS server. This embodiment is a case of a stream format that is sequentially transmitted from the SNS server to the EC server every time the user relation information is updated. In this case, the provider of the SNS site may make a contract with the provider of the EC site in advance to provide user relationship information in a stream format. In addition, user-related information may be provided for a fee.

  In the present embodiment, detailed description of the same configuration and operation as in the first embodiment will be omitted, and differences from the first embodiment will be described in detail.

  The structure of the SNS server 10 of a present Example is demonstrated. The information providing unit 111 of the SNS server 10 illustrated in FIG. 3 transmits user relationship information of all users to the EC server 20 every time there is a change in the user relationship information due to a new relationship between the users. .

  The configuration of the EC server 20 of this embodiment will be described. Each time the preference information estimation unit 215 of the EC server 20 receives user relationship information from the SNS server 10, the preference information estimation unit 215 performs an estimation process on the newly registered user preference information using the updated user relationship information, and performs estimation. The preference information 236 is updated.

  In addition, when there is a change in the human relationship of the existing user, the preference information estimation unit 215 uses the updated user relationship information to obtain the preference information for the existing user in the same manner as the newly registered user. The estimated preference information 236 or the preference information 235 stored in the storage unit 23 may be estimated and updated.

  In the first embodiment, when the EC server 20 requests user relationship information from the SNS server 10 and acquires the target user relationship information, the session between the EC server 20 and the SNS server 10 is disconnected. In the example, the state where the session between the EC server 20 and the SNS server 10 is connected is maintained.

  Also, as in the present embodiment, when the user relation information is received in the stream format and the estimation process is performed, the processing speed of the preference estimation becomes important. In the machine learning described in the first embodiment, it is preferable to use a large-scale real-time data processing platform such as Jubatus. However, the system to be used is not limited to Jubatus.

  In the present embodiment, since information can be received every time there is a change in the user relationship information, there is an advantage that the latest information is reflected in the estimation process. In addition, since the latest user relation information can be used not only for newly registered users but also for extracting preference information of existing users, it is possible to improve the accuracy of estimation of overall preference information.

  In this embodiment, the transmission method of the user relation information is in the stream format and the session connection between the apparatuses is maintained. However, every time the user relation information is transmitted, the session relation between the apparatuses is transmitted. It may be a push type that connects and disconnects and sequentially sends user relationship information.

  Further, this embodiment may be applied to the second embodiment. The EC server 22 shown in FIG. 15 combines the user relationship information 133 acquired from the SNS server 10 every time the user relationship information 133 is updated in the SNS server 10 with the user relationship information 233 held by the own device. Also in the second embodiment, the latest user relation information 133 can be reflected in the newly registered user preference estimation process.

  In the first embodiment, the case where the EC site cooperates with one SNS site has been described. However, the present embodiment is a case where the EC site cooperates with a plurality of SNS sites.

  In the present embodiment, detailed description of the same configuration and operation as in the first or second embodiment will be omitted, and differences from the first or second embodiment will be described in detail.

  The configuration of the communication system of the present embodiment will be described. FIG. 18 is a block diagram illustrating a configuration example of a communication system according to the present embodiment. As shown in FIG. 18, in the communication system of the present embodiment, SNS servers 10 a and 10 b with different SNS sites are connected to a network 50. The EC server 20 shown in FIG. 18 corresponds to the information processing apparatus 2 shown in FIG.

  FIG. 19 is a schematic diagram for explaining the configuration of the SNS server and EC server of this embodiment.

  The SNS servers 10a and 10b shown in FIGS. 18 and 19 have the same configuration as the SNS server 10 described in the first embodiment. In this embodiment, the user relationship information 133a is a social graph, but the user relationship information 133b. Is an interest graph. The interest graph here means a human relationship connected by a common hobby or event.

  Further, the SNS site a shown in FIG. 19 is a site opened by the SNS server 10a shown in FIG. 18 in the network 50, and the SNS site b is a site opened by the SNS server 10b in the network 50.

  Next, the operation of the EC server 20 in this embodiment will be described with reference to FIG. 13 and FIG.

  When the EC server 20 receives the user relationship information 133a and 133b from the plurality of SNS sites 10a and 10b in step 1004 of FIG. 13 for the newly registered user A, the EC server 20 adds the user relationship information 133a and 133b. The information is constructed as described in the second embodiment, and the preference information is estimated as described in the first embodiment in step 1006 shown in FIG.

  In this example, it is common to Example 2 in that a plurality of user relationship information is added, but if a plurality of types of graphs are included in the plurality of user relationship information, as follows: The difference is that the user relation information is weighted.

  As a weighting method, the service usage histories of the users associated with the user relationship information for each SNS site are compared, and if there is a correlation between the service usage histories of the users associated with the same user relationship information, the estimation is performed. It is possible to use a method that considers the user relationship information indicating a human connection closer to the target user's preference and increases the weight.

  A specific example of the weighting method will be described below.

The user relationship information 133a regarding the user A held by the SNS site “a” is assumed to be “user A is connected to each of the user B and the user C by a path”.
It is assumed that the user relationship information 133b regarding the user A held by the SNS site b is “user A is connected to each of the user D and the user E by a path”.

It is assumed that the following information is registered as service usage history at the EC site of users B to E. Here, for example, when the user B purchases a tennis ball, the service use history is described as “user name / purchased product”.
User B / tennis ball user C / anime DVD
User D / soccer shoe user E / soccer ball (omitted).

  The preference information estimation unit 215 refers to the service usage histories of the user B and the user C read from the user relationship information 133a, and is there a correlation between the “tennis ball” of the user B and the “anime DVD” of the user C? judge. In addition, the preference information estimation unit 215 refers to the service usage history of each of the user D and the user E read from the user relation information 133b, and correlates with the “soccer shoes” for the user D and the “soccer ball” for the user E. Determine if there is any. The preference information estimation unit 215 uses “soccer shoes” and “soccer balls” rather than the combination of “tennis balls” and “anime DVD” because “soccer” is common to “soccer shoes” and “soccer balls”. And the combination of “soccer ball” are judged to have a correlation.

  In this manner, the preference information estimation unit 215 derives “soccer” as the estimated preference information 236 of the user A, and determines that the user relationship information 133b of the SNS site b is human relationship information in which people with similar hobbies are collected, The user relationship information 133b is weighted higher than the user relationship information 133a.

  In the present embodiment, recommendation information can be provided to a newly registered user as in the second embodiment by combining a plurality of user relationship information acquired from a plurality of SNS sites.

  Moreover, although the tendency of the user relation information accumulated for each SNS site is divided into a social graph or an interest graph, the accuracy of the preference information estimation processing can be improved by performing the above-described weighting on the user relation information. This is because the human relationship information is roughly divided into a social graph and an interest graph, and the interest graph is closer to the user's preference than the social bluff and is suitable for estimating the preference information.

  In this embodiment, as an example of a method for weighting the user relationship information, the case where the type of the user relationship information is an interest graph or a social graph has been described. Not limited to cases. Further, the EC server 20 may be the EC server 22 described in the third embodiment. That is, the user relationship information to be combined is not limited to information acquired from a plurality of SNS sites, but may be user relationship information held by the EC server and user relationship information acquired from one or more SNS sites.

  In the information processing apparatus and the information processing method of the present embodiment, as specifically described in the first to fourth embodiments, the information recommendation cannot be received because there is almost no use history just after newly registered. Even the user can improve user convenience by estimating the preference information based on the user relationship information and performing information recommendation.

10, 10a, 10b SNS server 20, 22 EC server 21 Control unit 23 Storage unit 30 User terminal 50 Network

Claims (8)

A storage unit for storing preference information for each existing user who is an already registered user;
For a newly registered user who is a newly registered user, human relation information that is information indicating a person-to-person connection is acquired from another information processing apparatus via the network, and the new relation user is referred to and the new relation user is referred to. A control unit that identifies an existing user associated with a registered user, reads the preference information of the existing user from the storage unit, and estimates the preference information of the newly registered user based on the read preference information;
An information processing apparatus. The information processing apparatus according to claim 1,
Unlike the human relationship information held by the other information processing apparatus, human relationship information about the existing user is stored in the storage unit in advance.
The controller is
Using the combined human relationship information, which is human relationship information obtained by adding the human relationship information acquired from the other information processing apparatus and the human relationship information stored in the storage unit, the preference information of the newly registered user is obtained. Information processing apparatus to be estimated. The information processing apparatus according to claim 1 or 2,
The controller is
Each time the human relation information is updated in the other information processing apparatus, the human relation information received from the other information processing apparatus is used to estimate the preference information of the newly registered user. An information processing apparatus for updating preference information of a newly registered user. The information processing apparatus according to claim 1,
The controller is
When a plurality of the human relationship information is received from the plurality of other information processing apparatuses via the network, the new registration is performed using the combination human relationship information which is the human relationship information obtained by adding the plurality of human relationship information. An information processing apparatus that estimates user preference information. The information processing apparatus according to claim 2 or 4,
The controller is
An information processing apparatus that, when different types of human relationship information are included in the combined human relationship information, weights differently depending on the type of the human relationship information to be referred to. The information processing apparatus according to any one of claims 1 to 5,
The controller is
The use history of the newly registered user's own device is accumulated in the storage unit, and when the accumulated number of use histories is equal to or greater than a preset threshold, the specified existing user is identified when estimating the preference information of the newly registered user. An information processing apparatus that uses the use history of the newly registered user instead of the preference information. An information processing method executed by an information processing apparatus having a storage unit and a control unit,
The control unit prestores preference information for each existing user who is a registered user in the storage unit,
When the user is newly registered, the control unit transmits, to the other information processing apparatus, the human relationship information, which is information indicating a person-to-person connection, for the newly registered user who is a newly registered user via the network. Get from
The control unit identifies an existing user associated with the newly registered user with reference to the human relationship information,
The information processing method, wherein the control unit reads the specified preference information of an existing user from the storage unit, and estimates the preference information of the newly registered user based on the read preference information. In the information processing device,
A procedure for storing the preference information for each existing user who is already registered in the storage unit in advance;
A procedure for acquiring human relationship information, which is information indicating a person-to-person relationship, from another information processing apparatus via a network for a newly registered user who is a newly registered user when a user is newly registered; ,
Identifying an existing user associated with the newly registered user with reference to the human relationship information;
The program for performing the procedure which reads the said preference information of the specified existing user from the said memory | storage part, and estimates the preference information of the said newly registered user based on the read preference information.

Images