JP7369759B2 - Information processing system, information processing method, and program - Google Patents

Description

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

Conventionally, a credit score estimating unit that estimates the credit score of each user, which is the likelihood that the user will fulfill a debt in a credit transaction for the user, based on a first algorithm for deriving the credit score; a demand estimating unit that estimates the demand, which is the strength of the user's intention to perform credit transactions, based on a second algorithm for deriving the demand for each of the users; Information including the type or conditions of credit transactions selected according to the user group is distributed to user terminals used by users belonging to each user group in which the users are classified based on the combination of credit score and the demand. An information processing device has been disclosed (for example, see Patent Document 1).

Patent No. 6472915

However, with the above-mentioned conventional technology, it may not be possible to set an appropriate credit limit for the user.

The present invention takes such circumstances into consideration, and aims to provide an information processing system, a learning device, an information processing method, a program, and a learning method that can set a more appropriate credit limit for a user. Make it one of the objectives.

One aspect of the present invention includes an acquisition unit that acquires an attribute score based on a user's attribute and a credit score indicating the creditworthiness of the user; a determining unit that determines a group to which the user belongs; and a setting unit that configures a credit limit to be granted to the user based on the group determined by the determining unit, wherein each of the plurality of groups The information processing system is a group classified based on the attribute score, the credit score, the amount of debt, and the degree of bad debt.

According to one aspect of the present invention, a more appropriate credit limit can be set for a user.

It is a figure showing an example of a 1st zone and a 2nd zone. It is a figure which shows an example of a 3rd zone. 1 is a diagram showing an example of the configuration of an information processing system 1. FIG. 1 is a diagram illustrating an example of a functional configuration of a first information processing device 100. FIG. 3 is a diagram showing an example of the contents of user information 122. FIG. FIG. 3 is a diagram for explaining a process of acquiring an attribute score and a credit score. FIG. 3 is a diagram illustrating an example of user distribution for each combination of attribute score and trust score. FIG. 3 is a diagram showing the distribution of the amount of debts owed by users who have defaulted on their loans. It is a diagram showing the distribution of the total of receivables (revolving payments and cash advances). FIG. 2 is a diagram showing the distribution of loan default rates. 11 is a diagram showing the right diagram of FIG. 9 and the right diagram of FIG. 10. FIG. 2 is a flowchart illustrating an example of the flow of processing executed by the first information processing device 100. FIG. 2 is a diagram illustrating an example of a functional configuration of a second information processing device 200. FIG. 3 is a flowchart illustrating an example of the flow of processing executed by the second information processing device 200. FIG. FIG. 3 is a diagram for explaining zone settings. It is a figure showing an example of zone information 228A in which annual income is taken into account.

DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments of an information processing system, a learning device, an information processing method, a program, and a learning method of the present invention will be described with reference to the drawings.

[overview]
An information processing system according to one aspect of the present embodiment includes an acquisition unit that acquires an attribute score based on a user's attribute and a credit score indicating the creditworthiness of the user, and based on the attribute score and the credit score of the user, The device includes a determining unit that determines a group to which the user belongs among a plurality of groups, and a setting unit that configures a credit limit to be granted to the user based on the group determined by the determining unit. The plurality of groups are based on the attribute score, the credit score, the size of receivables (for example, the size of receivables from credit card services such as revolving payment, cash advance, and installment payments), and the degree of bad debt of each of the plurality of users. This is a group divided based on

The attribute score is a score obtained based on attribute information indicating the user's attributes including the user's gender, age, and occupation. The credit score is a score obtained based on credit information indicating the user's creditworthiness, including whether or not the user has borrowed money, the repayment status, and the amount of the loan.

Until now, as shown in Figure 1, when the user was divided into the first zone and the second zone based on the combination of the attribute score obtained from the user's attributes and the credit score obtained from the user's creditworthiness. was there. The first zone is, for example, an area where the attribute score and credit score are good, and the second zone is an area where the attribute score and credit score are not good. The first zone may be assumed to be a zone where the risk of default on loans is relatively low, and the second zone may be assumed to be a zone where the risk of default on loans is relatively high. In this embodiment, financing includes funding in the form of revolving credit card services, cash advances, installment payments, and the like. Details of the attribute score and credit score will be described later.

The information processing system of this embodiment found that a third zone exists in the second zone, as shown in FIG. The third zone is a zone for users who have a higher demand for loans and a lower risk of bad debt than users in the second zone. For example, the information processing system makes loan conditions for users in the third zone more relaxed than conditions for loans for users in the second zone. Relaxation of loan conditions means, for example, increasing the loan amount. Relaxation of loan conditions may include, in addition to the loan amount, relaxation of conditions that make the loan easier to use, such as loan period and loan interest. The above matters will be explained in detail below.

[Information processing system]
FIG. 3 is a diagram showing an example of the configuration of the information processing system 1. As shown in FIG. The information processing system 1 includes, for example, one or more user terminal devices U, a service server 10, a learning device 50, a first information processing device 100, and a second information processing device 200. These devices communicate with each other via the network NW. The network NW includes the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), a cellular network, and the like.

[Terminal device]
The terminal device U is a computer device, such as a smartphone or a tablet terminal, which is used by a user and has a calling function, a communication function, and the like.

[Service server]
The service server 10 is a device that provides services to the terminal device U. The service server 10 is, for example, a server device that provides a loan service (for example, revolving credit card service, cash advance, installment payment, etc.) to a user. The learning device 50 will be described later.

[First information processing device]
FIG. 4 is a diagram illustrating an example of the functional configuration of the first information processing device 100. The first information processing device 100 includes, for example, an acquisition unit 102, an attribute score acquisition unit 104, a credit score acquisition unit 106, a first statistical processing unit 108, and a second statistical processing unit 110 (an example of a “processing unit”). ), a zone setting section (another example of a "processing section") 112, a providing section 114, and a storage section 120. The acquisition unit 102, attribute score acquisition unit 104, credit score acquisition unit 106, first statistical processing unit 108, second statistical processing unit 110, zone setting unit 112, and provision unit 114 are, for example, CPUs (Central Processing Units), etc. This is realized by a hardware processor executing a program (software). Some or all of these components are hardware (circuit parts) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), and GPU (Graphics Processing Unit). (including circuitry), or may be realized by collaboration between software and hardware. The program may be stored in advance in a storage device (a storage device with a non-transitory storage medium) such as an HDD (Hard Disk Drive) or flash memory, or may be stored in a removable storage device such as a DVD or CD-ROM. It may be stored in a medium (non-transitory storage medium), and installed in the storage device by loading the storage medium into a drive device. The functions of each component will be described later. Note that in the first information processing device 100, some functional configurations (for example, the first statistical processing unit 108) may be omitted.

The storage unit 120 is realized by, for example, a ROM, a flash memory, an SD card, a RAM, an HDD, a register, or the like. The storage unit 120 stores, for example, user information 122, an attribute model 124, a trust model 126, first statistical information 128, second statistical information 130, and zone information 132.

The acquisition unit 102 acquires user information 122 from another device, and stores the acquired user information 122 in the storage unit 120. FIG. 5 is a diagram showing an example of the contents of the user information 122. The user information 122 is information in which user identification information (user ID) is associated with user attribute information and external credit information. The attribute information is information indicating user attributes such as the user's gender, age, and occupation, for example. The external credit information is, for example, information regarding external creditworthiness, such as whether or not the user has borrowed money from other companies, repayment status, and loan amount. The external credit information is, for example, information obtained from a credit information agency that provides external credit information.

The attribute score acquisition unit 104 acquires the user's attribute score based on the attribute information and the attribute model 124. The credit score acquisition unit 106 acquires the user's credit score based on external credit information and the credit model 126.

FIG. 6 is a diagram for explaining the process of acquiring attribute scores and credit scores. The attribute score acquisition unit 104 generates a user attribute vector from the attribute information. The user attribute vector is a distributed representation of user attribute features. The attribute score acquisition unit 104 inputs the user attribute vector into the attribute model 124 and acquires the output result (score) output by the attribute model 124.

When a user attribute vector is input, the attribute model 124 outputs an attribute score according to the user attribute vector. The attribute score is a score indicating the risk of bad debt.

The credit score acquisition unit 106 generates a user foreign credit vector from external credit information. The trust score acquisition unit 106 inputs the user foreign communication vector to the trust model 126 and acquires the output result (score) output by the trust model 126. When the user's foreign belief vector is input, the trust model 126 outputs a trust score corresponding to the user's foreign belief vector. A credit score is a score that indicates the risk of loan default.

The attribute model 124 and the trust model 126 (described later) are models using techniques such as logistics regression, support vector machine (SVM), decision tree, and neural network, for example. The attribute model 124 and the trust model 126 are models generated by a learning device 50 (not shown). The learning data and learning method used by the attribute model 124 and trust model 126 for learning will be described later.

Users are classified based on the combination of the above attribute score and credit score. FIG. 7 is a diagram illustrating an example of user distribution for each combination of attribute score and trust score. The left diagram in FIG. 7 is the distribution obtained by the first identification process, and the right diagram in FIG. 7 is the distribution obtained by the second identification process. In this embodiment, for example, the group to which the user belongs is determined using the results obtained by the second identification process.

The first specifying process is a process executed by the first statistical processing section 108, and the second specifying process is a process executed by the second statistical processing section 110. The diagrams on the left in FIGS. 7 to 10, which will be described later, are the first statistical information 128 generated in the first specifying process, and the diagrams on the right in FIGS. 7 to 10, which will be described later, are the second statistical information 128 generated in the second specifying process. This is statistical information 130. Hereinafter, the differences between the first specifying process and the second specifying process will be explained.

(Difference 1)
The attribute model that outputs attribute scores and the credit model that outputs credit scores are different. In the first identification process, a model (first attribute model, first credit model, not shown) that is generated by learning learning data in which users who have defaulted on loans are used as positive examples and users who have not defaulted on loans as negative examples is used. On the other hand, in the second identification process, the model ( Attribute model 124 and trust model 126) are used. Note that the information on the positive example and the information on the negative example may be reversed.

In the first specification process, an attribute score is obtained using a first attribute model for the first specification process, and a credit score is obtained using a first trust model for the first specification process. The first attribute model is a model obtained by learning the first attribute learning data by the learning device 50. The first attribute learning data includes positive example learning data in which information indicating that a loan has been defaulted and attribute information of a user who has defaulted on a loan are associated with each other, and information indicating that a loan has not been defaulted and attribute information of a user who has not defaulted on a loan. are learning data of negative examples that are associated with each other. The first attribute model outputs a relatively high score (score indicating that the probability of defaulting on a loan is high) when the attribute information of a user who has defaulted on a loan is input, and when the attribute information of a user who has not defaulted on a loan is input. , is a trained model that has been trained to output a relatively low score.

The first trust model is a model obtained by learning the first trust learning data by the learning device 50. The first credit learning data includes positive example learning data in which information indicating that a loan has been defaulted and external credit information of a user who has defaulted on a loan are associated with each other, and information indicating that there is no default on a loan and external credit information of a user who has not defaulted on a loan. This is negative example learning data in which credit information and credit information are associated with each other. The first credit model outputs a relatively high score when external credit information of a user who has defaulted on a loan is input, and outputs a relatively low score when external credit information of a user who has not defaulted on a loan is input. This is a trained model that has been trained as follows.

The attribute model 124 of the second identification process is a model obtained by the learning device 50 learning the second attribute learning data. The second attribute learning data includes learning data of a positive example in which information indicating that a user has defaulted on a loan with an amount of debt greater than or equal to a predetermined amount is associated with attribute information of a user who has defaulted on a loan with an amount of debt greater than or equal to a predetermined amount, and a predetermined amount. This is learning data of a negative example in which information indicating that the user has not defaulted on debts with a debt amount above a predetermined amount and attribute information of a user who has not defaulted on debts with a debt amount of more than a predetermined amount is associated with each other. The attribute model 124 outputs a relatively high score when the attribute information of a user who has defaulted on a debt with a debt amount equal to or more than a predetermined amount is input, and outputs the attribute information (or This is a trained model that has been trained to output a relatively low score when input user attribute information) is input.

The trust model 126 of the second identification process is a model obtained by the learning device 50 learning the second trust learning data. The second credit learning data includes positive example learning data in which information indicating that a user has defaulted on a debt with an amount of debt greater than a predetermined amount and external credit information of a user who has defaulted on a debt with an amount of debt greater than a predetermined amount are associated with each other; This is negative example learning data in which information indicating that there is no debt default with a debt amount greater than a predetermined amount and external credit information of a user who is not in debt debt with a debt amount greater than a predetermined amount are associated with each other. The credit model 126 outputs a relatively high score when the attribute information of a user who has defaulted on a debt with an amount of receivables greater than a predetermined amount is input, and outputs attribute information (or This is a trained model that has been trained to output a relatively low score when input user attribute information) is input.

Differences in the distributions shown in FIGS. 7 to 10 occur due to the different models that output attribute scores and credit scores as described above. For example, as will be described later, a group of users with a high total amount of receivables and a low default rate are emphasized.

(Difference 2)
In the first identification process and the second identification process, distributed users are grouped and statistical processing is performed to generate the left diagram in FIG. 7 and the right diagram in FIG. The grouping method used for statistical processing differs between processing and processing.

In the first identification process, scores were divided into predetermined intervals, and statistical processing was performed for each user who fell into the classification. For example, if the score is 0-1, the score is divided into 10 by 0.1, and statistical processing is performed on the user groups belonging to the divided divisions. As a result, as shown in the left diagram of FIG. 7, the number of users (the number of users with debts) for each statistical processing target group is displayed.

In the second identification process, the users are grouped so that the number of users in each statistical processing target group is equal. For example, in the distribution of users based on the combination of attribute scores and credit scores, multiple users are divided into unit groups so that the number of users included in the unit groups subject to statistical processing is the same or similar. For each unit group, the debt amount and the degree of default of each user included in the unit group are statistically processed.

For example, if there are 1000 users and they are divided into 100 statistical processing target groups, they are grouped so that each group includes 10 users. At this time, the users included in each statistical processing target group are balanced and grouped so that the combinations of attribute scores and credit scores are the same or close to each other. As a result, since there is a tendency for there to be many users with good attribute scores and trust scores, users with better attribute scores and trust scores are emphasized, as shown in the right diagram of FIG.

The left diagrams in each of FIGS. 8-10 described below are graphs obtained by the first specifying process, and the right diagrams in each of FIGS. 8-10 are graphs obtained by the second specifying process.

FIG. 8 is a diagram showing the distribution of debt amounts of users who have defaulted on loans. As shown in Figure 8, compared to the graph obtained from the first identification process, in the graph obtained from the second identification process, areas with large debts due to bad debts have expanded to areas with good attribute scores and credit scores. (The area of bad debts is expanding). From this, it can be understood that users who have defaulted on loans are widely distributed not only in areas where attribute scores and credit scores are poor.

FIG. 9 is a diagram showing the distribution of total receivables (revolving payments and cash advances). As shown in FIG. 9, compared to the graph obtained by the first identification process, in the graph obtained by the second identification process, areas with large claims extend to areas with good attribute scores and credit scores.

FIG. 10 is a diagram showing the distribution of loan loss rates. As shown in Figure 10, compared to the graph obtained by the first identification process, the graph obtained by the second identification process shows that areas with a high loan loss rate have spread to areas with a good credit score, but the attribute score has not spread to good areas.

FIG. 11 is a diagram showing the right diagram of FIG. 9 and the right diagram of FIG. 10. The area where the area AR1 in the left diagram of FIG. 11 and the area AR2 in the right diagram of FIG. 11 overlap is estimated to be an area where frame expansion is possible. The area AR1 is an area where the total amount of receivables is large, and is estimated to be an area to which users who have a high demand for revolving payments or cash advances belong. The area AR2 is an area where the rate of bad debts is not high or low, and is estimated to be an area to which users who have a low risk of bad debts even if revolving payment or cash advance services are provided belong to the area.

As described above, the zone setting unit 112 extracts the area AR1 and the area AR2 based on the result of the statistical processing performed by the second specifying process, and the zone setting unit 112 extracts the area AR1 and the area AR2, and the third Identify the zone. The zone setting unit 112 generates zone information 132 regarding the specified third zone. The zone information 132 is information indicating a combination of attribute scores and credit scores of users classified into the third zone.

The zone setting unit 112 is configured to perform a process based on the result of statistical processing of the debt amount of the plurality of users based on the combination of the attribute score and the credit score, and the combination of the attribute score and the credit score. This is an example of a "processing unit that extracts the group (for example, the third group) based on the result of statistical processing of the degree of bad debt." In addition, the zone setting unit 112 determines that “based on the results of statistical processing of the amount of debt and the degree of debt default of the plurality of users, the debt amount is equal to or more than a first threshold value, and the degree of debt debt is equal to or less than a second threshold value” This is an example of a processing unit that sets a group (for example, a third group) including users having a combination of an attribute score and the credit score.

The providing unit 114 provides the attribute model 124, the trust model 126, and the zone information 132 to the second information processing device 200.

(flowchart)
FIG. 12 is a flowchart illustrating an example of the flow of processing executed by the first information processing device 100. The order of this process may be changed, or some processes (for example, "S12") may be omitted.

First, the acquisition unit 102 acquires user information 122 (S10). Next, the first statistical processing unit 108 generates first statistical information 128 based on the user information 122, the attribute score and the trust score obtained from the first attribute model and the first trust model (S12). Next, the second statistical processing unit 110 generates second statistical information 130 based on the attribute score and trust score obtained from the user information 122, the attribute model 124, and the trust model 126 (S14). Next, the zone setting unit 112 generates zone information 132 (S16). For example, the zone setting unit 112 generates zone information 132 based on the second statistical information 130 described above. This completes the processing of one routine in this flowchart. Note that the zone setting unit 112 may set the zone based on the first statistical information 128.

As described above, the first information processing device 100 can extract the third group that includes users who have a higher demand for loans and a lower loan default rate. Note that the above processing may be performed at predetermined intervals. For example, the first information processing device 100 may generate the second statistical information 130 at predetermined intervals, and update the zone information 132 based on the latest generated second statistical information 130. As a result, the third zone is set according to the latest information.

[Second information processing device]
FIG. 13 is a diagram illustrating an example of the functional configuration of the second information processing device 200. The second information processing device 200 includes, for example, an information management section 202, a score acquisition section 204, a classification section 206, a setting section 208, an information provision section 210, and a storage section 220. The information management section 202, score acquisition section 204, classification section 206, setting section 208, and information provision section 210 are realized by, for example, a hardware processor such as a CPU executing a program (software). Some or all of these components may be realized by hardware (including circuitry) such as LSI, ASIC, FPGA, or GPU, or may be realized by collaboration between software and hardware. Good too. The program may be stored in advance in a storage device such as an HDD or flash memory (storage device equipped with a non-transitory storage medium), or may be stored in a removable storage medium (non-transitory storage medium) such as a DVD or CD-ROM. The software may be installed in the storage device by attaching the storage medium to the drive device. The functions of each component will be described later.

The storage unit 220 is realized by, for example, a ROM, flash memory, SD card, RAM, HDD, register, or the like. The storage unit 220 stores, for example, user information 222, an attribute model 224, a trust model 226, and zone information 228.

The user information 222 includes attribute information and external credit information of the user for whom a credit limit is to be set. The attribute model 224 is the attribute model 124 provided by the first information processing device 100. The trust model 226 is the trust model 126 provided by the first information processing device 100. The zone information 228 is the zone information 132 provided by the first information processing device 100.

Information management unit 202 acquires user information. The information management unit 202 acquires user information from the service server 10, for example. The score acquisition unit 204 acquires an attribute score based on the attribute information of the user information and the attribute model 224. The score acquisition unit 204 acquires a credit score based on the external credit information of the user information and the credit model 226.

The classification unit 206 determines to which zone the user belongs based on the attribute score and the trust score. The setting unit 208 sets a credit limit for the user based on the determination result of the classification unit 206. The information providing unit 210 provides the service server 10 with information on the set credit limit.

(flowchart)
FIG. 14 is a flowchart illustrating an example of the flow of processing executed by the second information processing device 200. First, the information management unit 202 acquires a request regarding credit limit determination and user information from the service server 10 (S50). Next, the score acquisition unit 204 acquires the attribute score and the credit score (S52). Next, the classification unit 206 determines which zone the user belongs to based on the acquired attribute score and trust score (S54).

Next, the setting unit 208 sets a credit limit for the user based on the classification result (S56). Next, the information providing unit 210 provides information on the set credit limit to the service server 10 (S58). Thereby, the service server 10 can determine a credit limit based on the loan-related request obtained from the user's terminal device U, and can provide a loan to the user within the range of the determined credit limit.

FIG. 15 is a diagram for explaining zone settings. The classification unit 206 refers to the zone information 228 shown in FIG. 15 to determine a zone corresponding to a position based on the attribute score and the credit score. For example, if the position P based on the attribute score and the trust score is in the third zone, the classification unit 206 determines that the user belongs to the third zone. The setting unit 208 makes the credit limit for users in the third group larger than the credit limit for users in the second zone. In other words, the setting unit 208 sets a credit limit for a user who belongs to a group whose debt amount is equal to or higher than a first threshold and whose degree of defaulted debt is equal to or lower than a second threshold out of a plurality of groups, when the degree of defaulted debt is equal to the second threshold. be larger than the credit limit for users belonging to groups exceeding .

If the target user who has set the first credit limit in the past belongs to a group in which the amount of debt is equal to or higher than the first threshold and the degree of loan default is equal to or lower than the second threshold, the setting unit 208 sets the target user who has set the first credit limit in the past to Set up a larger second line of credit. For example, in the past, the third zone may not have been found and the user may have been classified into the first zone or the second zone. In this case, even a user who belongs to the third zone is classified as belonging to the second zone, and the loan conditions are set according to the conditions based on the second zone.

In the present embodiment, as described above, when the third zone is found, the setting unit 208 selects a user who has been determined to belong to the second zone in the past and who satisfies the conditions for belonging to the third zone. By classifying the user as belonging to the third zone, a more appropriate credit limit can be set for the user belonging to the third zone.

Note that the credit limit set for the user, the user's attribute score, and credit score described in FIGS. 14 and 15 above may be updated at predetermined intervals. For example, when a user applies for issuance of a credit card, the user's attribute score and credit score are acquired (scoring is performed), and a credit limit is set. After that, scoring will be re-performed every six months or other arbitrary period, the user's attribute score and credit score will be updated to the latest, and a credit limit will be set according to the updated user's attribute score and credit score. may be done. In this way, an appropriate credit limit is set by performing scoring every arbitrary period.

[Modified example]
In a modified example, the user's annual income may be taken into account. The user's annual income may be the user's individual annual income, or may include the annual household income or the annual income of another person who shares the household finances with the user. FIG. 16 is a diagram showing an example of zone information 228A in which annual income is taken into account. The zone information 228A is information in which zones 1 to 3 are divided by group and annual income. Annual income is highest in order of A-F. Group 1-50 (50, but any number may be used) is a group in which 50 combinations of attribute scores and credit scores are classified. For example, a group with a small combination of attribute score and credit score is classified as "1", and as the combination becomes larger, the group is classified into 2, 3, 4, . . . 50 in order. For example, a combination with the largest combination of attribute score and credit score is classified into group 50.

Note that in the above grouping, the groups may be divided based on a score obtained by multiplying an attribute score or a credit score by a coefficient. For example, emphasis may be placed on the score of the model that outputs a more accurate score between the attribute model 224 and the trust model 226. For example, if the accuracy of the credit score output by the credit model 226 is higher than the accuracy of the attribute score output by the attribute model 224, the credit score or attribute score is adjusted so that the weight of the credit score is increased, and the adjusted They may be grouped based on their scores. This allows more suitable grouping.

As shown in the figure, for example, the group with a small number is the first zone, and as the annual income increases, the first zone expands to the group with a large number. As shown in the figure, for example, the group with the higher number is the second zone, and as the annual income decreases, the second zone expands to the group with the lower number. In the illustrated example, for example, areas with annual incomes E and F do not fall under the second zone. An area that does not belong to the first zone or the second zone is the third zone. Additionally, loan limits are set for each combination of annual income and group in each zone.

The second information processing device 200 obtains an attribute score and a trust score based on user information. The second information processing device 200 determines which group the user belongs to based on the acquired attribute score and trust score. The second information processing device 200 determines which zone the user falls into among the first to third zones based on the determined group and the user's annual income. The second information processing device 200 determines a loan limit associated with the determined zone.

As described above, the second information processing device 200 can appropriately set a credit limit for the user by taking into account the user's annual income.

According to the information processing system 1 of the embodiment described above, the user determines the group to which the user belongs from among the plurality of groups classified based on the attribute score, credit score, debt amount, and degree of bad debt. By setting a user's credit limit based on the selected group, a more appropriate credit limit can be set for the user.

Also, some of the functions executed by the first information processing device 100 or the second information processing device 200 described above may be executed by another device (for example, the service server 10, etc.). Further, the information stored in the first information processing device 100 or the second information processing device 200 may be stored in another device.

Although the mode for implementing the present invention has been described above using embodiments, the present invention is not limited to these embodiments in any way, and various modifications and substitutions can be made without departing from the gist of the present invention. can be added.

1‥Information processing system, 10‥Service server, 50‥Learning device, 100‥First information processing device, 102‥Acquisition unit, 104‥Attribute score acquisition unit, 106‥Credit score acquisition unit, 108‥First statistical processing unit , 110... Second statistical processing unit, 112... Zone setting unit, 114... Providing unit, 120... Storage unit, 122... User information, 124... Attribute model, 126... Trust model, 128... First statistical information, 130... 2 statistical information, 132... zone information, 200... second information processing device, 202... information management section, 204... score acquisition section, 206... classification section, 208... setting section, 210... information provision section, 220... storage section, 222‥User information, 224‥Attribute model, 226‥Trust model, 228‥Zone information

Claims (10)

an acquisition unit that acquires an attribute score based on a user's attributes and a credit score indicating the user's creditworthiness;
a determining unit that determines a group to which the user belongs among a plurality of groups based on the attribute score and the credit score of the user;
a setting unit that sets a credit limit to be granted to the user based on the group determined by the determination unit;
a processing unit that sets the group ;
The plurality of groups are groups classified based on the attribute score, the credit score, the debt amount , and the degree of bad debt of each of the plurality of users,
The plurality of users include a plurality of users with different combinations of the attribute score and the credit score,
The processing unit includes:
a result of statistical processing of the debt amount of the plurality of users based on the combination of the attribute score and the credit score;
Extracting the group based on the result of statistical processing of the degree of bad debt of the plurality of users based on the combination of the attribute score and the credit score,
In the user distribution based on the combination of the attribute score and the credit score, the plurality of users are divided into unit groups such that the number of users included in the unit groups to be statistically processed is the same or similar, and the Statistically processing the debt amount and the degree of debt default of each user included in the unit group group for each divided unit group group;
Information processing system. The attribute score is a score obtained based on attribute information indicating the user's attributes including the user's gender, age, and occupation,
The credit score is a score obtained based on credit information indicating the user's creditworthiness, including whether or not the user has borrowed money, repayment status, and loan amount.
The information processing system according to claim 1. The acquisition unit includes:
Obtaining the attribute score output by the attribute model into which the attribute information has been input,
obtaining the credit score output by the credit model into which the credit score information has been input;
The attribute model takes as a positive example information indicating the attributes of a user who has a debt amount of more than a predetermined amount and who has defaulted on a loan, and uses information indicating the attributes of a user who has a debt amount of more than a predetermined amount and has not defaulted on a loan as a negative example. This is a model that has learned the training data.
The information processing system according to claim 2. The acquisition unit includes:
Obtaining the attribute score output by the attribute model into which the attribute information has been input,
obtaining the credit score output by the credit model into which the credit score information has been input;
The credit model takes as a positive example information indicating the creditworthiness of a user who has a debt amount greater than a predetermined amount and who has defaulted on a loan, and uses information indicating the creditworthiness of a user who has a debt amount greater than a predetermined amount and has not defaulted on a loan as a negative example. This is a model that has learned the training data.
The information processing system according to claim 2 or 3. The processing unit is configured to calculate the attribute score in which the debt amount is equal to or higher than a first threshold and the degree of debt default is equal to or lower than a second threshold based on the result of statistical processing of the debt amount and the degree of debt default of the plurality of users. Set up a group that includes users in combination with a credit score,
The information processing system according to any one of claims 1 to 4. The setting unit is configured to set a credit limit to a user belonging to a group in which the debt amount is equal to or higher than a first threshold and the degree of defaulted debt is equal to or lower than a second threshold among the plurality of groups, and the setting unit sets a credit limit to a user who belongs to a group in which the debt amount is equal to or higher than a first threshold and the degree of defaulted debt is equal to or lower than a second threshold. be larger than the credit limit for users belonging to groups exceeding
The information processing system according to any one of claims 1 to 5. The setting section includes:
If a target user who has set a first credit limit in the past belongs to a group in which the debt amount is equal to or greater than the first threshold and the degree of bad debt is equal to or less than the second threshold, the target user has a first credit limit larger than the first credit limit. 2 Set up a credit limit,
The information processing system according to any one of claims 1 to 6. The setting unit further sets a credit limit for the user based on the user's annual income in addition to the group.
The information processing system according to any one of claims 1 to 7. The computer is
obtaining an attribute score based on the user's attributes and a credit score indicating the user's creditworthiness;
determining a group to which the user belongs among a plurality of groups based on the attribute score and the credit score of the user;
setting a credit limit to be granted to the user based on the determined group;
set the group ,
The plurality of groups are groups classified based on the attribute score, the credit score, the debt amount , and the degree of bad debt of each of the plurality of users,
The plurality of users include a plurality of users with different combinations of the attribute score and the credit score,
a result of statistical processing of the debt amount of the plurality of users based on the combination of the attribute score and the credit score;
Extracting the group based on the result of statistical processing of the degree of bad debt of the plurality of users based on the combination of the attribute score and the credit score,
In the distribution of users based on the combination of the attribute score and the credit score, the plurality of users are divided into unit groups such that the number of users included in the unit groups to be subjected to statistical processing is the same or similar, and the Statistically processing the debt amount and the degree of debt default of each user included in the unit group group for each divided unit group group;
Information processing method. to the computer,
obtaining an attribute score based on the user's attributes and a credit score indicating the user's creditworthiness;
determining a group to which the user belongs among a plurality of groups based on the attribute score and the credit score of the user;
setting a credit limit to be granted to the user based on the determined group;
set the group ;
The plurality of groups are groups classified based on the attribute score, the credit score, the debt amount , and the degree of bad debt of each of the plurality of users,
The plurality of users include a plurality of users with different combinations of the attribute score and the credit score,
a result of statistical processing of the debt amount of the plurality of users based on the combination of the attribute score and the credit score;
extracting the group based on the result of statistical processing of the degree of bad debt of the plurality of users based on the combination of the attribute score and the credit score;
In the distribution of users based on the combination of the attribute score and the credit score, the plurality of users are divided into unit groups such that the number of users included in the unit groups to be subjected to statistical processing is the same or similar, and the Statistically processing the debt amount and the degree of debt default of each user included in the unit group group for each divided unit group group;
program.

Images