JP2021136033A - Information processing device, information processing method and program - Google Patents

Abstract

To provide an information processing device, an information processing method and a program capable of properly evaluating a user.SOLUTION: In a user evaluation system in which the information processing device, an analysis server, a user terminal, an affiliated server and a financial institution server are mutually connected for communication via a network such as Internet, a first user DB 141 of a server 1 which is the information processing device stores transaction information indicating a transaction history of a first user and action history information indicating an action history of the first user on the network. A control section 11 executes a program P to acquire identification information provided to a second user due to a session to a Web site, acquires action history information of the second user on the basis of the acquired identification information, and evaluates the second user on the basis of the transaction information and the action history information of the first user and the action history information of the second user.SELECTED DRAWING: Figure 2

Description

The present invention relates to an information processing device, an information processing method and a program.

When conducting commercial transactions, financial transactions, etc. with customers, the credit of the customers is evaluated. Along with this, various methods for evaluating customer credit have been proposed.

For example, in Patent Document 1, the asset information of a second user having an action history similar to that of the first user is estimated based on the asset information of the user and the action history of the user on the network, and the second user A determination device or the like for determining the creditworthiness of a user is disclosed.

Japanese Unexamined Patent Publication No. 2018-1100117

In the invention according to Patent Document 1, another user (first user) similar to the user to be evaluated (second user) is extracted from the action history on the network, and the search engine is used as the action history. The search query entered in and the articles posted to the SNS (Social Networking Service) will be used. However, in order to collect these action histories, it is necessary to register user information (terminals used by users, SNS accounts, etc.) in advance, which is not easy.

In one aspect, it is an object of the present invention to provide an information processing device or the like capable of appropriately evaluating a user.

The information processing device according to one aspect includes a storage unit that stores transaction information indicating the transaction history of the first user and action history information indicating the action history of the first user on the network, and a session to a website. The first acquisition unit that acquires the identification information given to the second user, the second acquisition unit that acquires the action history information of the second user based on the identification information, and the transaction of the first user. It is characterized by including an evaluation unit that evaluates the second user based on the information and the action history information and the action history information of the second user.

In one aspect, the user can be favorably evaluated.

It is a schematic diagram which shows the configuration example of the user evaluation system. It is a block diagram which shows the configuration example of a server. It is explanatory drawing which shows an example of the record layout of the action history DB and the customer DB. It is explanatory drawing which shows an example of the record layout of the 1st user DB and the determination table. It is a timing chart which shows the procedure of the analysis process of the action history. It is explanatory drawing about the analysis processing of the action history information. It is explanatory drawing about a keyword extraction process. It is a timing chart which shows the procedure of a user evaluation process. It is explanatory drawing which shows the processing example of the user evaluation system. It is a flowchart which shows the processing procedure which a server executes. It is explanatory drawing which shows the application example of the user evaluation system. It is a block diagram which shows the configuration example of the server which concerns on modification 2. It is a flowchart which shows the procedure of the generation process of a determination model.

Hereinafter, the present invention will be described in detail with reference to the drawings showing the embodiments thereof.
(Embodiment)
FIG. 1 is a schematic diagram showing a configuration example of a user evaluation system. In the present embodiment, based on the financial transaction history of a large number of first users (other users) who are customers of a financial institution and the behavior history of each first user on the Internet, any second user A user evaluation system that evaluates creditworthiness on credit will be described. The user evaluation system includes an information processing device 1, an analysis server 2, a user terminal 3, an affiliated server 4, and a financial institution server 5. The devices are communicated and connected to each other via a network N such as the Internet.

The information processing device 1 is an information processing device capable of transmitting and receiving various types of information processing and information, and is, for example, a server computer, a personal computer, or the like. In the present embodiment, it is assumed that the information processing device 1 is a server computer, and in the following, it will be read as server 1 for the sake of brevity. Preferably, the server 1 shows transaction information (for example, transaction history of a financial product) indicating a transaction history of a first user who is a customer of a financial institution (bank or the like) and an action history of the first user on the network N. Based on the action history information (for example, a word in the website viewed by the first user), the creditworthiness of any second user is determined (evaluated).

In the present embodiment, the user who is the basis of the credit rating evaluation is called a "first user", and the user to be evaluated is called a "second user" to distinguish them. For example, the first user and the second user may be the same person. When the two are not distinguished, they are simply referred to as "users". Further, the user is not limited to an individual, but may be a corporation.

Further, in the present embodiment, the creditworthiness of the second user is determined from the transaction history of financial products and the like, but the present embodiment is not limited to this. For example, the server 1 may determine the creditworthiness from the purchase history (history of commercial transactions) of products at a physical store, an EC site, or the like. That is, the transaction information is not limited to the history of financial transactions.

Further, in the present embodiment, the creditworthiness in credit is mainly determined, but the present embodiment is not limited to this. For example, the server 1 may determine the purchase probability of the second user purchasing the product, as will be described later. The server 1 may evaluate the user based on the transaction information and the action history information, and the scale of the evaluation is not particularly limited.

The analysis server 2 is a server computer that analyzes the user's behavior history on the network N, and is a server computer that analyzes the Web pages viewed by the user on a plurality of affiliated websites (hereinafter referred to as "affiliated sites"). be. In the present embodiment, the analysis server 2 collects the data of the browsing page and analyzes the action history information by using a method called DMP (Data Management Platform). The analysis server 2 stores the analyzed action history information in the action history DB (Database) 201.

DMP is a platform for integrating and managing data stored in various Web servers and optimizing advertisement distribution. In the present embodiment, DMP is applied to credit evaluation, and the credit rating is determined using the analysis result by the analysis server 2, that is, the user's behavior history information.

The user terminal 3 is a terminal device operated by the user, for example, a personal computer, a smartphone, a tablet terminal, or the like. The analysis server 2 analyzes the Web page of the affiliated site accessed by the user terminal 3.

The affiliated server 4 is a server computer of an operator who operates an affiliated site, and is a server computer that functions as a Web server. Examples of affiliated sites include EC (Electronic Commerce) sites, news article sites, search sites, etc., but the contents are not particularly limited.

The financial institution server 5 is a server computer of a financial institution used by the first user as a customer, and provides the customer transaction information stored in the customer DB 501 to the server 1. Further, in the present embodiment, the financial institution server 5 is also affiliated with the analysis server 2 to provide a website of the financial institution (for example, an electronic money transfer site, a credit statement confirmation site, etc.). In the following description, for convenience, it is referred to as a "financial institution site" to distinguish it from other affiliated sites.

In the present embodiment, the analysis server 2 centrally manages the identification information given to the user at each affiliated site, collects the data of the Web page viewed by the user associated with the identification information, and analyzes the action history. ..

Generally, in a website, when a session is received from a user, a session ID is assigned to each session and data is managed. The ID is stored in the user terminal 3 as a value of Cookie data, and the Web server (affiliated server 4) identifies the terminal of the session source based on the ID.

However, since each website assigns an ID separately, even if you try to link session data on multiple websites for a certain user, the session data on one website will be any session on another website. Cannot identify whether it corresponds to the data.

Therefore, for example, a program called a DMP tag is embedded in the affiliated sites AA, BB, CC, etc. that are affiliated with the analysis server 2. When the user terminal 3 accesses a certain affiliated site AA, the user terminal 3 executes a DMP tag to access the analysis server 2 and receives issuance of an ID common to the affiliated sites AA, BB, CC, and so on. When accessing other affiliated sites BB, CC ..., The user terminal 3 sends the ID to perform a session, so that the session data at each affiliated site can be linked.

In the following description, the ID given by the analysis server 2 is referred to as a "common ID".

The above-mentioned linking method is an example, and the method is not limited to the one called DMP. The analysis server 2 may collect data based on the identification information (for example, the above-mentioned common ID) given to the user at a certain affiliated site, and the specific method thereof is not particularly limited. Further, the data collection method is not limited to the method using Cookie, and may be a method using access information such as an IP address and a user agent.

A DMP tag is also embedded in the financial institution site, and when the user terminal 3 of the first user accesses the financial institution site, the common ID given to the first user is sent to the analysis server 2. As a result, the transaction information of the first user held by the financial institution can be associated with the behavior history information of the first user analyzed by the analysis server 2.

The server 1 determines the creditworthiness of the second user based on the transaction information and the action history information of the first user. Then, the server 1 provides the determination result (evaluation result) to the affiliated server 4 of the business operator who conducts the transaction with the second user. For example, when a business operator accepts a product purchase application from a second user via an affiliated site, such as an EC site, the server 1 has application information indicating the purchased product, price, etc., and a common ID of the second user. Is received from the affiliated server 4 and the credit rating output request is accepted. The server 1 determines the credit rating in response to the request and outputs the determination result.

In the following, the explanation will be given mainly assuming the case of purchasing products on the EC site, but the services (transactions) provided on the affiliated sites are not limited to the purchase of products, such as financial loans, insurance, real estate sales, etc. It may be. Further, it is not necessary for the server 1 to provide the credit rating determination result to all the business operators of the affiliated sites (affiliated server 4), and it may be provided only to some of the business operators. Further, the affiliated site that collects the behavior history information and the affiliated site of the business operator that conducts transactions with the second user may be different.

Further, the server 1 that performs the credit rating determination process, the analysis server 2 that analyzes the user's behavior history, and the financial institution server 5 that holds the user's transaction information will be described as separate server computers. The device may be the same server computer.

Further, in the following description, the creditworthiness of the user is provided to the business operator of the affiliated site (affiliated server 4), but of course, it may be provided to the financial institution (financial institution server 5).

FIG. 2 is a block diagram showing a configuration example of the server 1. The server 1 includes a control unit 11, a main storage unit 12, a communication unit 13, and an auxiliary storage unit 14.
The control unit 11 has one or more CPUs (Central Processing Units), MPUs (Micro-Processing Units), GPUs (Graphics Processing Units), and other arithmetic processing units, and stores the program P stored in the auxiliary storage unit 14. By reading and executing, various information processing, control processing, etc. are performed. The main storage unit 12 is a temporary storage area for SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), flash memory, etc., and temporarily stores data necessary for the control unit 11 to execute arithmetic processing. Remember. The communication unit 13 is a communication module for performing processing related to communication, and transmits / receives information to / from the outside.

The auxiliary storage unit 14 is a non-volatile storage area such as a large-capacity memory or a hard disk, and stores a program P and other data necessary for the control unit 11 to execute processing. Further, the auxiliary storage unit 14 stores the first user DB 141 and the determination table 142. The first user DB 141 is a database that stores transaction information and action history information of the first user. The server 1 associates the transaction information of the first user who is a customer of the financial institution and the action history information obtained by analyzing the Web page of the affiliated site browsed by the first user with each other in the first user DB 141. Remember. The determination table 142 is a table used for determining the creditworthiness of the second user, and is a table for storing a keyword (reference character string) to be referred to in the determination process.

The auxiliary storage unit 14 may be an external storage device connected to the server 1. Further, the server 1 may be a multi-computer composed of a plurality of computers, or may be a virtual machine virtually constructed by software.

Further, in the present embodiment, the server 1 is not limited to the above configuration, and may include, for example, an input unit that accepts operation input, a display unit that displays an image, and the like. Further, the server 1 is provided with a reading unit for reading a portable storage medium 1a such as a CD (Compact Disk) -ROM, a DVD (Digital Versatile Disc) -ROM, and reads and executes a program P from the portable storage medium 1a. You can do it. Alternatively, the server 1 may read the program P from the semiconductor memory 1b.

FIG. 3 is an explanatory diagram showing an example of the record layout of the action history DB 201 and the customer DB 501.
The action history DB 201 includes a common ID string and a word string. The common ID column stores the common ID given to the first user. The word string is associated with the common ID and stores the word (character string) described in the Web page viewed by the first user.

The customer DB 501 includes a customer ID column, a name column, a common ID column, a deposit balance column, a card balance column, and a transaction column. The customer ID column is an ID for identifying a first user who is a customer of a financial institution. The name column, common ID column, deposit balance column, card balance column, and transaction column are associated with the customer ID, respectively, of the name of the first user, the common ID given to the first user, the deposit balance, and the credit card. It stores the available balance and the transaction history of financial products.

FIG. 4 is an explanatory diagram showing an example of the record layout of the first user DB 141 and the determination table 142.
The first user DB 141 includes a data number string, a deposit balance column, a card balance column, a transaction column, and a word string. The data number string stores the data number for associating the transaction information and the action history information. The deposit balance column, the card balance column, the transaction column, and the word string are associated with the data number to store the deposit balance of the first user, the credit card usage balance, the transaction history of the financial product, and the word on the browsing page, respectively. doing. The data stored in the deposit balance column, the card balance column, and the transaction column correspond to the transaction information, and the data stored in the word string corresponds to the action history information.

The determination table 142 includes a transaction type column and a keyword column. The transaction type column stores the type of financial transaction. The keyword string stores the keywords to be referred to when determining the creditworthiness in association with the transaction type.

FIG. 5 is a timing chart showing the procedure of the action history analysis process. The behavior history analysis process using the common ID will be described with reference to FIG.
The user terminal 3 of the first user accesses the affiliated site in response to the operation input from the first user (step S101). The affiliated server 4 responds to the access from the user terminal 3 and transmits the data (HTML file) of the Web page of the affiliated site (step S102). The data transmitted in step S102 includes a DMP tag for assigning a common ID.

The user terminal 3 calls the DMP tag, redirects to the analysis server 2, and transmits the data of the Web page of the affiliated site accessed in step S101 to the analysis server 2 (step S103). For example, the user terminal 3 transmits text data such as a title and a text of a Web page. When the common ID is not assigned to the user terminal 3, the analysis server 2 assigns the common ID and stores the data of the Web page in association with the common ID (step S104). The user terminal 3 stores the common ID given by the analysis server 2.

The user terminal 3 accesses the affiliated site in response to the operation input from the first user (step S105). The affiliated site accessed in step S105 may be the same as or different from the affiliated site accessed in step S101. The affiliated server 4 transmits the data of the affiliated site including the DMP tag in response to the access from the user terminal 3 (step S106). The user terminal 3 calls the DMP tag and transmits the stored common ID and the data of the Web page accessed in step S105 to the analysis server 2 (step S107). The analysis server 2 stores the data of the Web page in association with the common ID.

The analysis server 2 analyzes the behavior history of the first user from the data of the Web page collected based on the common ID by, for example, batch processing (step S108). For example, the analysis server 2 extracts words included in the Web page viewed by the user, such as the title of the browsing page and meta keywords (keywords designated as important words by the tags of the HTML file).

FIG. 6 is an explanatory diagram relating to the analysis process of the action history information. As described above, when the user terminal 3 accesses the affiliated site, the analysis server 2 assigns a common ID by calling the DMP tag. When the user terminal 3 accesses the affiliated site, the user terminal 3 transmits the common ID and the data of the browsing page to the analysis server 2.

The analysis server 2 extracts words described in the page, such as a title and a meta keyword, from the browsing page collected based on the common ID. The analysis server 2 stores (stores) the extracted words in the database in association with the common ID assigned to the user terminal 3.

In the present embodiment, words in the browsing page are extracted as action history information, but the present embodiment is not limited to this. For example, the analysis server 2 analyzes the session data between the user terminal 3 and the affiliated server 4, and the URL of the page of the affiliated site viewed by the first user, the referrer (the first user links from the external site to the affiliated site). (Link source page when accessing), advertisements clicked by the first user, etc. may be extracted. Alternatively, when the affiliated site is a search site, the search query entered by the first user may be analyzed. As described above, the action history information is not limited to the words in the page viewed by the first user.

Further, in the above, as the action history information, words, URLs, referrers, etc. in the page viewed by the user are illustrated, but the present embodiment is not limited to this. For example, when the page viewed by the user is the product posting page of the EC site, the information of the posted product may be extracted as the action history information. As described above, the action history information may be data indicating the user's action on the network N, and the content thereof is not particularly limited.

Further, in the above, the character string of the browsing page is extracted for each word, but the character string may be extracted for each phrase, sentence, or the like.

Further, in the above, the text data of the browsing page is directly sent from the user terminal 3 to the analysis server 2, but for example, when the user terminal 3 accesses the affiliated site, the analysis server 2 only sends the URL and session ID of the browsing page. You may record it and later access the affiliated server 4 to acquire the text data of the page viewed by the first user from the affiliated server 4.

The explanation will be continued by returning to FIG. The user terminal 3 accesses the financial institution site according to the operation input from the first user (step S109). The financial institution server 5 responds to the access from the user terminal 3 and transmits the data of the Web page of the financial institution site (step S110). The data transmitted in step S110 includes a DMP tag.

The user terminal 3 calls the DMP tag and transmits the common ID to the financial institution server 5 (step S111). The financial institution server 5 stores a common ID in association with the customer ID. By the above processing, the behavior history information of the first user is accumulated in the analysis server 2, and the first user who is a customer is associated with the common ID in the financial institution server 5.

In the above, it was explained that the first user has already been given a common ID when accessing the financial institution site, but if the affiliated site accessed for the first time is the financial institution site, the financial institution site is accessed. It goes without saying that a common ID is assigned at the same timing.

Based on the common ID acquired in step S111, the financial institution server 5 requests the analysis server 2 to output the behavior history information of the first user corresponding to the common ID (step S112). The analysis server 2 transmits the behavior history information of the first user to the financial institution server 5 (step S113). Specifically, as described above, the analysis server 2 transmits the word data extracted from the browsing page of the first user to the financial institution server 5. The financial institution server 5 transmits the action history information received from the analysis server 2 and the corresponding transaction information of the first user to the server 1 (step S114).

As described above, the transaction information is data indicating the transaction history of the first user at a financial institution, and in addition to the deposit balance of the first user, the available balance of a credit card, etc., the type of transaction (for example, the type of financial product). ) It is data indicating whether or not there is a contract or cancellation for each (see FIG. 3). The financial institution server 5 provides the transaction information and the corresponding behavior history information of the first user to the server 1 as a series of data sets as a basis for evaluating the second user. The server 1 stores the transaction information of the first user and the action history information associated with the common ID of the first user in the first user DB 141 in association with each other.

When the transaction information is transmitted from the financial institution server 5, anonymized data is transmitted so that an individual is not identified from the transaction history. For example, the financial institution server 5 does not transmit personally identifiable data such as a common ID and the name of the first user, but only transmits transaction details. In addition, only a part of the various data included in the transaction information is transmitted to the server 1 so that the individual is not identified from the transaction history. Alternatively, when the financial institution server 5 transmits a numerical value such as a deposit balance, it transmits a rounded numerical value instead of a raw numerical value. This prevents the first user from being identified from the transaction information.

In the present embodiment, anonymized data is transmitted to the server 1 so that an individual cannot be identified, but it goes without saying that raw data including personal information may be provided to the server 1. ..

Further, in the present embodiment, the financial institution server 5 collectively transmits the transaction information and the action history information of the first user to the server 1, but the present embodiment is not limited to this. For example, the financial institution server 5 may provide the transaction information of the first user to the analysis server 2, and the analysis server 2 may transmit the transaction information and the action history information to the server 1. Alternatively, when the server 1 can acquire the common ID of the first user, the server 1 requests the analysis server 2 and the financial institution server 5 to output the action history information and the transaction information based on the common ID, and from each of them. You may get the data. As described above, the acquisition route when the server 1 acquires the transaction information and the action history information is not particularly limited.

The server 1 extracts a keyword (reference character string) to be referred to when evaluating the second user from the transaction information and the action history information of the first user (step S115). The server 1 stores the extracted keywords in the determination table 142.

FIG. 7 is an explanatory diagram relating to the keyword extraction process. Based on FIG. 7, the processing content when extracting the keyword in step S115 will be described.
As described above, the behavior history information of the first user and the transaction information are associated with each other based on the common ID given to the first user. The server 1 extracts the keyword used when determining the creditworthiness of the second user from the action history information based on the transaction information associated with the action history information based on the common ID.

For example, as shown in the lower left of FIG. 7, the server 1 extracts words in the page viewed by the first user as keywords and stores them in the determination table 142. Specifically, the server 1 extracts keywords for each transaction type (for example, the type of financial product) according to the presence or absence of the transaction, and stores the keywords in the determination table 142. For example, when the "delinquency" of the "card loan" is "yes" for a first user, the server 1 associates the words extracted from the page viewed by the first user with the "card loan" and "delinquency". And store it in the determination table 142.

The server 1 extracts keywords for each transaction type as described above. When determining the creditworthiness of the second user, the server 1 compares the words in the page viewed by the second user with the keywords, and determines the creditworthiness of the user when conducting various different transactions.

FIG. 8 is a timing chart showing the procedure of the user evaluation process. Based on FIG. 8, the processing content when determining the creditworthiness of the second user will be described.
The user terminal 3 of the second user accesses an affiliated site (for example, an EC site) for conducting commercial transactions, financial transactions, etc. according to the operation input from the second user (step S201). The affiliated server 4 responds to the access from the user terminal 3 and transmits the data of the Web page of the affiliated site (step S202). The data transmitted in step S202 includes a DMP tag in addition to information related to transactions (for example, product information). The user terminal 3 calls the DMP tag and transmits the common ID to the analysis server 2 (step S203).

The user terminal 3 transmits the application information for applying for a transaction from the second user to the business operator of the affiliated site to the affiliated server 4 according to the operation input from the second user (step S204). When the affiliated site is an EC site, the user terminal 3 transmits a purchase application for the product. In step S204, the user terminal 3 transmits the common ID to the affiliated server 4 in addition to the application information.

When the application information is acquired from the user terminal 3, the affiliated server 4 transmits the application information and the common ID to the server 1 and requests the output of the credit rating (step S205). When the credit rating output request is received, the server 1 transmits the common ID acquired from the affiliated server 4 to the analysis server 2 and requests the output of the behavior history information of the second user corresponding to the common ID (step S206). .. In response to the request from the server 1, the analysis server 2 outputs the behavior history information of the second user (step S207).

The server 1 determines (evaluates) the creditworthiness of the second user based on the behavior history information of the second user acquired from the analysis server 2 and the transaction information and behavior history information of the first user, and the determination result (evaluation result). ) Is output to the affiliated server 4 (step S208). The affiliated server 4 determines whether or not the transaction is possible based on the credit rating output from the server 1, and notifies the user terminal 3 of the determination result (step S209).

For example, the server 1 compares the words in the page viewed by the second user with the keywords of various transactions stored in the determination table 142, and determines the creditworthiness for each transaction type. For example, in the case of "mortgage", the server 1 collates the word in the browsing page with each keyword such as "contract", "cancellation", and "bankruptcy" related to "mortgage", and the second user makes a mortgage. Judge the creditworthiness when conducting a transaction related to.

The specific method for determining (calculating) the creditworthiness is not particularly limited, but for example, the creditworthiness is calculated by determining positive and negative points for each keyword and accumulating these points. The server 1 calculates the overall credit rating of the second user by summing the credit ratings determined for each transaction type.

The above determination method is an example, and the credit rating may be determined by referring to, for example, the URL of the page viewed by the second user, the referrer, and the like in addition to the words in the browsing page.

Further, for example, the server 1 may weight the credit rating by referring to the deposit balance of the first user whose browsing page is the same as or similar to that of the second user, the available balance of the credit card, and the like. The similarity of browsing pages may be determined, for example, according to the frequency of appearance of common words (keywords).

Further, for example, the server 1 may determine the credit rating for each period by referring to the action history information for each predetermined period (for example, every month). This makes it possible to output the change in the creditworthiness of the second user.

Further, for example, the server 1 may output the breakdown of the credit rating determination factors (for example, the score for each transaction type) based on the score of the action history information which is the basis of the credit rating determination (calculation). As a result, the external business operator can grasp the factors for determining the creditworthiness.

Further, in the above, the credit rating is determined based on the rule, but as in the modified example 2 described later, machine learning is performed based on the action history information and transaction information accumulated in the customer DB 501 and the action history DB 201, and the action history information is input. Then, a model (for example, a neural network) that outputs the credit rating may be constructed. As a result, the system administrator does not need to set a rule, and the user's behavior history information can be input to the model to determine the creditworthiness.

FIG. 9 is an explanatory diagram showing a processing example of the user evaluation system. FIG. 9 illustrates a hypothetical example of the processing described in the timing chart of FIG.
As described above, the business operator related to the affiliated server 4 conducts a transaction with the second user via the affiliated site. In the example of FIG. 9, assuming that the affiliated site is an EC site, it is assumed that the second user pays the purchase price to the business operator later by means of payment such as a credit card.

When the application information is acquired from the user terminal 3, the affiliated server 4 transmits the application information indicating the details of the transaction applied by the second user and the common ID of the second user to the server 1, and outputs the credit rating. Request. The application information includes, for example, information such as a product desired by the second user and a purchase price, as well as information on the payment method (credit card, etc.) of the second user.

When the application information is acquired, the server 1 transfers the common ID to the analysis server 2 and acquires the behavior history information of the second user. Then, the server 1 refers to the determination table 142 and determines the creditworthiness of the second user from the action history information.

In the present embodiment, the creditworthiness of the second user is determined from the financial transaction history of the first user, but the server 1 is a commercial transaction that the business operator of each affiliated site has conducted with the first user in the past. The creditworthiness of the second user may be determined with reference to the above history. Specifically, when the server 1 acquires the common ID and application information of the first user from each business operator (affiliated server 4), the acquired application information is stored in the first user DB 141 as transaction information. .. Further, the server 1 requests the output of the behavior history information of the first user based on the common ID, acquires the behavior history information from the analysis server 2, and associates it with the transaction information of the first user to the first user DB 141. Remember.

When the server 1 receives the output request of the creditworthiness of the second user from one business operator, the server 1 receives the transaction information on the financial transaction of the first user acquired from the financial institution server 5, and the second user acquired from the affiliated server 4. 1 The credit rating is determined by referring to the transaction information on the user's commercial transaction. In the example of FIG. 9, when the credit rating output request is received from the "EC site XX", the server 1 determines the credit rating by referring to the transaction history of the first user at the "EC site XX". The server 1 may use the transaction history of another affiliated site different from the affiliated site to which the second user applied for the transaction for determining the creditworthiness of the second user. By referring to the transaction history of the first user who has used the same affiliated site in the past, the creditworthiness of the second user can be suitably determined based on the characteristics of the user of the affiliated site.

The method of determining the creditworthiness including the transaction history on the affiliated site is not particularly limited, but for example, the first user who has a similar transaction tendency is identified from the purchased product, the purchase amount, etc. applied by the second user, and is similar to the second user. If the first user has a non-payment transaction history, a method such as determining that there is a high possibility of default (credit loss) and lowering the credit rating can be considered. In this way, the server 1 may determine the creditworthiness by referring to the business transaction history of the first user that has been performed in the past with the business operator of the affiliated site.

Further, for example, the server 1 may lower the creditworthiness of the second user when the application information from the same second user is acquired from a plurality of affiliated sites (EC sites). Specifically, when the transactions (deferred payment) indicated by the application information acquired from each affiliated site occur at the same time, the server 1 lowers the creditworthiness of the second user. As a result, the creditworthiness can be appropriately determined in consideration of the current transaction status (application status) on the plurality of affiliated sites.

The server 1 outputs the credit rating determination result to the affiliated server 4. It is preferable that the credit rating is output as a numerical value (for example, a value of 0 to 1), but the result of ranking in the form of rank A, rank B, rank C, for example may be output. The affiliated server 4 determines whether or not to approve the application based on the credit rating output from the server 1, and notifies the user terminal 3 of the determination result.

In the above, the creditworthiness of the second user in commercial transactions and financial transactions is determined, but the present embodiment is not limited to this. For example, the server 1 estimates (evaluates) the purchase probability that the second user purchases the product on the EC site based on the behavior history information and the transaction information of the first user, and provides the estimation result to the EC site. good. In this case, for example, the server 1 determines the similarity between the behavior history information of the first user and the second user, and extracts the first user similar to the second user. The server 1 estimates the purchasing power of the second user and the products of interest from the extracted transaction information of the first user, and outputs the estimation result to the affiliated server 4. As described above, the server 1 only needs to be able to evaluate the second user based on the behavior history information, the transaction information, and the like, and the evaluation scale is not limited to the creditworthiness in the transaction.

FIG. 10 is a flowchart showing a processing procedure executed by the server 1. The processing contents executed by the server 1 will be described with reference to FIG.
The control unit 11 of the server 1 sets the application information indicating the details of the transaction applied to the business operator by the second user and the common ID (identification information) given to the second user by the session to the affiliated site. , Obtained from the affiliated server 4 of the business operator (step S11).

The control unit 11 transmits the common ID acquired in step S11 to the analysis server 2 and requests the output of the action history information indicating the action history of the second user on the network N (step S12). For example, the control unit 11 requests the output of a URL, a referrer, a click advertisement, or the like, in addition to the word (character string) described on the page of the affiliated site viewed by the second user. The control unit 11 acquires the action history information from the analysis server 2 (step S13).

The control unit 11 determines the creditworthiness of the second user based on the behavior history information of the second user acquired in step S13 and the behavior history information and transaction information of the first user who is a customer of the financial institution (step S14). ). Specifically, the determination table 142 stores keywords (reference character strings) for each transaction type, and the control unit 11 sets the words and keywords in the page viewed by the user for each transaction type. Compare and judge creditworthiness. Further, the control unit 11 determines the creditworthiness by referring to the transaction history (transaction information) of each transaction type. In addition, the control unit 11 may determine the credit rating from the URL, referrer, click advertisement, and the like. Further, the control unit 11 may determine the creditworthiness by referring to the commercial transaction history (transaction information) of the first user acquired from each business operator. The control unit 11 outputs the determination result to the affiliated server 4 (step S15), and ends a series of processes.

In the above, the transaction information and the action history information of each user are linked based on the common ID (DMP tag), but the present embodiment is not limited to this, and can be a user's identifier. It may be linked based on the identification information of. Other identification information is, for example, access information of the user terminal 3 that can be acquired when accessing a website, specifically, an IP address, a user agent, a time stamp, and the like.

For example, the analysis server 2 stores the access information when the user accesses the affiliated site in the action history DB 201 in association with the action history information at the time of analyzing the action history information of each user. When the analysis server 2 receives the output request of the behavior history information of the second user from the server 1 and / or the financial institution server 5, the analysis server 2 acquires the access information in addition to the common ID or in place of the common ID. .. The analysis server 2 compares the acquired access information with the access information of each user stored in the action history DB 201, estimates the same or similar users, and outputs the action history information of the user. In this way, it suffices if the behavior history information and transaction information can be linked based on the identification information related to the user that can be acquired when accessing the website (affiliated site), and the identification information that serves as the reference is the common ID. Not limited to.

From the above, according to the present embodiment, the behavior history information of the second user is acquired based on the identification information given according to the session to the affiliated site (website), and the behavior history information of the first user and the behavior history information The second user can be appropriately evaluated (preferably credit evaluation) with reference to the transaction information.

Further, according to the present embodiment, by acquiring the action history information from the analysis server 2 (management device) that manages the identification information (common ID) of the second user at each affiliated site, the behavior history information is acquired on the network N. It is possible to preferably grasp the behavior history of the second user in.

Further, according to the present embodiment, the behavior history of the first user from the analysis server 2 is acquired or given according to the session to the financial institution site (first website). Information can be preferably obtained.

Further, according to the present embodiment, by comparing the word (character string) described on the page of the affiliated site viewed by the second user with the keyword (reference character string), the second user can be more preferably used. Can be evaluated.

Further, according to the present embodiment, by referring to the keyword and the transaction history for each type of transaction, it is possible to more preferably perform the user evaluation by combining the action history information and the transaction information.

Further, according to the present embodiment, the evaluation result (for example, credit rating) of the second user can be provided to the external business operator.

Further, according to the present embodiment, the second user can be more preferably evaluated by referring to the transaction contents (application information) that the first user has made with the business operator in the past.

(Modification example 1)
In the above-described embodiment, a case where the creditworthiness of the second user is mainly determined (evaluated) has been described. On the other hand, the evaluation scale of the second user does not have to be the credit rating.

FIG. 11 is an explanatory diagram showing an application example of the user evaluation system. FIG. 11 schematically shows a case where this system is applied to various uses. Specifically, it is divided into target users (individuals or corporations), target services ("financial", "non-financial", etc.), and uses ("credit", "promotion", "fraud detection", "others"). , The keyword (behavior history information) that serves as a reference when evaluating the target second user is conceptually illustrated. In the above-described embodiment, the case of determining the creditworthiness on credit (column of "credit" in FIG. 11) has been described, but as shown in FIG. 11, the server 1 has "promotion", "fraud detection", and "fraud detection". The second user may be evaluated for purposes such as "others".

"Sales promotion" is a case of promoting the sale of an object such as a product to a second user. The "product" referred to here is not limited to an actual product, but may include a financial product, an insurance product, and the like. Further, the "object" is not limited to a product, but may include the provision of a service.

For example, the server 1 stores the transaction history (whether or not there is a transaction, etc.) of various objects by the first user in the first user DB 141 as transaction information. Then, the server 1 acquires the keyword of the Web page viewed by each first user from the analysis server 2 and stores it in the determination table 142 in association with the transaction history of the object. When the server 1 receives the output request of the behavior history information of the second user from the affiliated server 4, the server 1 acquires the behavior history information (keyword) of the second user from the analysis server 2 and stores the keyword in the determination table 142. Identify a first user who has the same or similar keywords as the second user. Then, the server 1 determines an object for which sales to the second user should be promoted based on the transaction history of the first user, and outputs the determination result to the affiliated server 4.

Explaining with reference to FIG. 11, for example, in the case of promoting the sales of the object (financial product) related to "finance" to the "individual" user ("grasp of necessary products"), the server 1 is Keywords such as "introduction to investment" and "new construction" are stored in the judgment table 142. The server 1 compares the keywords of the Web page viewed by the second user with these keywords, and determines the object (financial product) for which sales to the second user should be promoted.

"Fraud detection" is a case where the second user detects (determines) the possibility of committing fraudulent acts such as false declaration at the time of transaction. The server 1 may determine the possibility of fraudulent activity (hereinafter referred to as “fraudulent risk”) from the behavior history information of the user on the network N.

For example, the server 1 stores the presence or absence of fraudulent activity of the first user at the time of transaction as the transaction information of the first user in the first user DB 141. Then, the server 1 acquires the behavior history information of the first user who has committed fraudulent activity from the analysis server 2, and the type of fraudulent activity (false declaration, multiple contracts with false user name from the same terminal, borrowing). The keyword is stored in the determination table 142 for each (illegal use, etc.). When the server 1 receives the output request of the behavior history information of the second user from the affiliated server 4, the server 1 acquires the behavior history information of the second user from the analysis server 2 and compares it with the keyword stored in the determination table 142. Then, the fraud risk by the second user is determined. The server 1 outputs the determination result to the affiliated server 4 and warns.

As "others", it is assumed that the second user (customer) is used for prediction of separation, a call to the second user (sales by telephone, e-mail, etc.), prediction of moving, and the like. As a case used for prediction of separation, for example, when the second user is a customer of a financial institution, the server 1 determines a keyword related to the first user (user who canceled the account, etc.) who has separated from the financial institution in the determination table 142. It is stored in the server, and the possibility that the second user is separated is determined. Further, as a case of using the server 1, for example, when the second user is a customer of a financial institution and the second user makes a call to introduce a financial product or the like to the second user, each user makes a call. The time zone for accessing the Web page (browser) is extracted from the access information, and the time zone for calling the second user is determined. Further, as a case used for moving prediction, for example, in the server 1, a change in the access information (IP address, user agent, etc.) of the second user (customer) at a plurality of time points causes a change in the residence of the second user (moving). ) Is determined.

In this way, the server 1 only needs to be able to evaluate the second user based on the behavior history information of the second user, and can be used for various evaluation purposes such as sales promotion, fraud detection, etc. in addition to determination of creditworthiness in credit. is assumed. As a result, in addition to credit evaluation, this system can be applied to various measures such as narrowing down sales promotion targets, strengthening identity verification, implementing separation prevention campaigns, and creating customer call lists.

The server 1 may combine these evaluation methods. For example, when the second user accesses the affiliated site, the server 1 first determines the risk of fraud by the second user. If it is determined that there is no fraud risk, the server 1 then determines the creditworthiness of the second user. When it is determined that there is no problem in the credit rating (for example, the credit rating is equal to or higher than a predetermined threshold value), the server 1 further determines an object for which sales should be promoted. The server 1 outputs each determination result to the affiliated server 4. In this case, for example, the server 1 may store the keywords for each determination in a plurality of determination tables 142, 142, 142 ..., And use the corresponding determination table 142 when performing each determination. In this way, each evaluation method may be combined. Further, for example, the server 1 may allow the affiliated business operator, which is the output destination of the credit rating, to select one of these evaluation methods.

Since it is the same as the above-described embodiment except that the evaluation method of the second user is different, detailed description of the flowchart and the like will be omitted in this modification.

From the above, according to the present modification 1, the second user can be evaluated by various methods.

(Modification 2)
In the above-described embodiment, the mode in which the second user is evaluated on a rule basis has been described. On the other hand, a machine learning model may be constructed using the behavior history information and transaction information of the first user as teacher data, and the second user may be evaluated using the model.

FIG. 12 is a block diagram showing a configuration example of the server 1 according to the modification 2. The auxiliary storage unit 14 of the server 1 according to this modification stores the determination model 50 instead of the determination table 142. The determination model 50 is a machine learning model in which the behavior history information and transaction information of the first user have been learned, and when the behavior history information of the second user is input, the evaluation value of the second user regarding credit and the like is determined. It is a model. The determination model 50 is expected to be used as a program module that constitutes a part of artificial intelligence software.

For example, the determination model 50 is a neural network generated by deep learning using the behavior history information and transaction information of the first user as teacher data. In addition to the neural network, the determination model 50 may be a model based on other learning algorithms such as a decision tree, a random forest, and an SVM (Support Vector Machine).

The evaluation value output from the determination model 50 is, for example, the creditworthiness of the second user (for example, the probability of delinquency in postpaid settlement). The evaluation value is not limited to the credit rating, and may be the purchase probability, fraud risk probability, separation probability, etc. of the promotional object described in the first modification. Further, the evaluation value may be a continuous regression estimation value (for example, a value of 0 to 1) or a classification value (for example, a rank in which the credit rating is graded).

The server 1 generates the determination model 50 by learning the behavior history information of the first user collected based on the common ID and the transaction information of the first user. Specifically, the server 1 assigns the correct answer value of the credit rating to each first user according to the action history information of each first user. For example, the server 1 refers to the first user DB 141 (see FIG. 4), adjusts the credit rating according to the presence or absence of a contract, delinquency, cancellation, etc. with each first user, and assigns a correct answer value. The server 1 creates teacher data by labeling the behavior history information of each first user with the correct answer value of the credit rating given to each first user. The server 1 generates the determination model 50 using the created teacher data.

The correct answer value of the credit rating (evaluation value) is not automatically given by the server 1, but may be given manually.

The server 1 inputs the behavior history information for the teacher into the determination model 50, and acquires the credit rating as the output from the determination model 50. The server 1 compares the credit rating output from the determination model 50 with the correct answer value, and updates parameters such as weights between neurons so that the two are close to each other. The server 1 learns teacher data for each first user, and finally generates a determination model 50 with optimized parameters.

Here, the action history information input to the determination model 50 in the present embodiment includes access information that can be acquired when accessing the affiliated site, in addition to the keywords extracted from the affiliated site browsed by the first user. The access information is log data when the user terminal 3 accesses the affiliated site, such as an IP address and a user agent. The server 1 gives access information other than the keyword to the determination model 50 as teacher data for learning.

In particular, the access information to be learned in the present embodiment includes the domain of the Web page viewed by the user, the language setting in the user terminal 3, the user agent, and the IP address. According to the development of the inventor of the present application, the accuracy of determining the credit rating can be improved by including these access information in the feature amount.

When determining the creditworthiness of the second user, the server 1 determines the creditworthiness of the second user by inputting the behavior history information (including access information) of the second user acquired from the affiliated server 4 into the determination model 50. do. The server 1 outputs the determined credit rating to the affiliated server 4.

FIG. 13 is a flowchart showing the procedure of the generation process of the determination model 50. Based on FIG. 13, the processing content when the determination model 50 is generated by machine learning will be described.
The control unit 11 of the server 1 acquires the action history information and the transaction information of each first user from the first user DB 141 (step S21). The action history information includes access information in addition to the keywords extracted from the affiliated site (website) accessed by the first user. The access information is log data when the user terminal 3 accesses the affiliated site, and includes the domain of the Web page viewed (accessed) by the user, the language setting in the user terminal 3, the IP address, and the user agent.

The control unit 11 assigns a correct answer value of the creditworthiness of the first user based on the transaction information of the first user (step S22). The correct answer value may be automatically assigned by the server 1 or may be manually assigned.

The control unit 11 generates a determination model 50 that determines the credit rating when the action history information is input, based on the action history information of the first user and the correct answer value of the credit rating given in step S202 (step S23). For example, the control unit 11 generates a neural network as a determination model 50. The control unit 11 acquires the creditworthiness of the first user from the determination model 50 by inputting the behavior history information of the first user into the determination model 50. The control unit 11 optimizes parameters such as weights between neurons so that the acquired credit rating is close to the correct answer value. As a result, the control unit 11 generates the determination model 50. The control unit 11 ends a series of processes.

Since it is the same as the above-described embodiment except that the credit rating of the second user is determined by using the determination model 50, detailed description of the flowchart (see FIG. 10) at the time of credit rating determination is omitted in this modification. do.

Access information such as the domain, language setting, IP address, and user agent of the browsing page can also be used when evaluating the second user (determining the creditworthiness) on a rule basis as in the above-described embodiment. good. For example, when the server 1 stores the keyword of the Web page viewed by the first user in the determination table 142, the server 1 simultaneously stores the domain, language setting, IP address, and user agent of the browsing page of the first user. Then, the server 1 acquires each address information together with the keyword as the action history information of the second user, and determines the credit rating by comparing with the keyword and the address information stored in the determination table 142. As described above, the server 1 only needs to be able to evaluate the second user based on the behavior history information including each access information, and the evaluation method based on the access information is not limited to the evaluation by the machine learning model.

From the above, according to the present modification 2, the second user can be evaluated using the machine learning model.

The embodiments disclosed this time should be considered as exemplary in all respects and not restrictive. The scope of the present invention is indicated by the scope of claims, not the above-mentioned meaning, and is intended to include all modifications within the meaning and scope equivalent to the scope of claims.

1 Server (information processing device)
11 Control unit 12 Main storage unit 13 Communication unit 14 Auxiliary storage unit P program 2 Analysis server 3 User terminal 4 Affiliated server 5 Financial institution server

Claims (15)

A storage unit that stores transaction information indicating the transaction history of the first user and action history information indicating the action history of the first user on the network.
The first acquisition unit that acquires the identification information given to the second user by the session to the website, and
A second acquisition unit that acquires the behavior history information of the second user based on the identification information, and
An information processing device including an evaluation unit that evaluates the second user based on the transaction information and the action history information of the first user and the action history information of the second user. The information processing device according to claim 1, wherein the evaluation unit determines the creditworthiness of the second user. The transaction information includes the transaction history of a predetermined object by the first user.
The information processing apparatus according to claim 1 or 2, wherein the evaluation unit determines the object for which sales to the second user should be promoted. The transaction information includes the presence or absence of fraudulent activity by the first user at the time of transaction.
The information processing device according to any one of claims 1 to 3, wherein the evaluation unit determines the possibility of fraudulent activity by the second user at the time of transaction. The evaluation unit uses the behavior history of the second user in a model trained to determine the evaluation value of the user when the behavior history information is input based on the transaction information and the behavior history information of the first user. The information processing apparatus according to any one of claims 1 to 4, wherein the evaluation value of the second user is determined by inputting information. A request unit that requests the output of the action history information based on the identification information acquired by the first acquisition unit is provided for the management device that manages the identification information given by each of the plurality of the websites.
The information processing apparatus according to any one of claims 1 to 5, wherein the second acquisition unit acquires the behavior history information indicating the behavior history of the second user on the plurality of websites. .. The transaction information includes information indicating a financial transaction history of the first user who is a customer of a financial institution.
In response to a session from the first user to the website of the financial institution, the transaction information of the first user and the first user on the plurality of websites associated with the identification information of the first user. It is equipped with a third acquisition unit that acquires the above-mentioned action history information.
The information processing device according to claim 6, wherein the storage unit stores the transaction information and the action history information of the first user in association with each other. The action history information is data indicating a character string described on a page of the website.
An extraction unit for extracting a reference character string to be referred to when evaluating the second user from the behavior history information of the first user based on the transaction information of the first user is provided.
Any of claims 1 to 7, wherein the evaluation unit evaluates the second user by comparing the character string indicated by the action history information of the second user with the reference character string. The information processing apparatus according to item 1. The extraction unit extracts the reference character string for each type of transaction.
The information processing apparatus according to claim 8, wherein the evaluation unit evaluates the second user by comparing the character string with the reference character string for each type of transaction. The action history information further includes the browsing domain, language setting, user agent, or IP address of the second user.
The information processing device according to claim 8 or 9, wherein the evaluation unit evaluates the second user based on the character string and the browsing domain, language setting, user agent, or IP address. The first acquisition unit acquires the identification information of the second user from a business operator who conducts a transaction with the second user via the website.
The information processing apparatus according to any one of claims 1 to 10, further comprising an output unit that outputs the evaluation result of the second user to the business operator. The transaction information includes information indicating a history of commercial transactions performed by the first user with the business operator.
The first acquisition unit acquires the identification information and application information indicating the contents of the transaction applied to the business operator by the second user.
The eleventh claim is characterized in that the evaluation unit evaluates the second user based on the transaction information and the action history information of the first user and the application information and the action history information of the second user. The information processing device described in. Acquires the identification information given to the user by the session to the website,
Based on the identification information, the behavior history information indicating the behavior history of the user on the network is acquired, and the behavior history information is acquired.
The computer performs a process of evaluating the user based on the action history information of the user, the action history information of another user stored in the storage unit, and transaction information indicating the transaction history of the other user. An information processing method characterized by execution. Acquires the identification information given to the user by the session to the website,
Based on the identification information, the behavior history information indicating the behavior history of the user on the network is acquired, and the behavior history information is acquired.
The computer is subjected to a process of evaluating the user based on the action history information of the user, the action history information of another user stored in the storage unit, and transaction information indicating the transaction history of the other user. A program characterized by being executed. A storage unit that stores transaction information indicating the transaction history of the first user and action history information indicating the action history of the first user on the network.
When accessing the website, the first acquisition unit that acquires the identification information related to the second user who accessed the website, and the first acquisition unit.
A second acquisition unit that acquires the behavior history information of the second user based on the identification information, and
An information processing device including an evaluation unit that evaluates the second user based on the transaction information and the action history information of the first user and the action history information of the second user.

Images