TW202424876A - Information processing system, information processing method and program product which is applied to a load review based on the credit score and loyalty score estimated for the target user - Google Patents

Abstract

A subject of the present invention is a method to improve user evaluation in the business ecosystem. This information processing system includes: a credit score estimation unit (23), which estimates the user's credit score based on the attribute data group; a loyalty score calculation unit (25), which calculates the user's loyalty score based on the service utilization status in the business ecosystem; and a comprehensive score calculation unit (26), which calculates the comprehensive score of the target user based on the credit score and loyalty score estimated for the target user. The system also includes a gift decision means that determines the gift content based on the comprehensive score and credit score of the target user.

Description

Information processing system, information processing method and program product

This disclosure relates to the techniques required to calculate user scores.

Previously, a determination device comprising: a user information acquisition unit that acquires action information representing the user's actions; and a credit determination unit that determines the creditworthiness of the user regarding the future ability to repay financing based on the action information has been proposed (see patent document 1).

In addition, a loan review device that performs a loan review on a user based on input review application information and personal confirmation data has been proposed previously (see Patent Document 2). Furthermore, in the past, an information processing method has been proposed that accepts a request for a loan from a user, conducts a review based on at least one of user attribute information and user action information to determine loan conditions, notifies the user of the determined loan conditions, and, upon receiving a notification indicating that the user accepts the loan conditions, lends money to the user in accordance with the loan conditions (see Patent Document 3). [Prior Art Document] [Patent Document]

[Patent Document 1] Japanese Patent Publication No. 2021-174039 [Patent Document 2] Japanese Patent Publication No. 2020-003869 [Patent Document 3] Japanese Patent Publication No. 2020-102007

[The problem that the invention wants to solve]

Previously, technologies have been proposed for calculating user scores that represent user creditworthiness based on user profile data. However, in a business ecosystem (economic circle) where a variety of different services are gathered, there is still room for improvement in user evaluation based solely on credit scores.

This disclosure is based on the above-mentioned problems and aims to improve the user evaluation methods in the business ecosystem. [Methods to solve the problem]

One example of the present disclosure is an information processing system, which has: a credit score estimation means for estimating a credit score to be set for a user based on an attribute data group related to the user; a loyalty score calculation means for calculating the user's loyalty score with respect to a predetermined business ecosystem based on service utilization conditions performed by previous users in the predetermined business ecosystem; and a comprehensive score calculation means for calculating a comprehensive score of a target user based on a previous credit score estimated for the target user and a previous loyalty score calculated for the target user.

The present disclosure can be defined as a method executed by an information processing device, system, or computer, or a program executed by a computer. In addition, the present disclosure can also be defined as recording such a program on a recording medium readable by a computer or other device, machine, etc. Here, the so-called recording medium readable by a computer, etc. refers to a recording medium that stores information such as data or programs by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer, etc. [Effect of the invention]

According to this disclosure, the user evaluation method in the business ecosystem can be improved.

The following will describe the implementation forms of the information processing device, method and program product involved in the present disclosure based on the drawings. However, the implementation forms described below are only illustrative implementation forms, and the information processing device, method and program product involved in the present disclosure are not limited to the specific configuration described below. During implementation, the specific configuration can be appropriately adopted according to the implementation mode, and various improvements or modifications can be made. The present invention can appropriately adopt at least a part of the configuration of each of the implementation forms and modifications described below.

In this embodiment, the technology involved in this disclosure is described in order to provide a user of a predetermined business ecosystem that can use multiple online services with a gift or promote the use of services. However, the technology involved in this disclosure can be widely used in the technology required to determine the points of users, and the application of this disclosure is not limited to the examples shown in the embodiment.

<System Configuration> Figure 1 is a schematic diagram of the configuration of the information processing system described in this embodiment. In the information processing system described in this embodiment, the information processing device 1 and one or more service providing systems 5 are connected to communicate with each other. The user is a user of the service belonging to the predetermined business ecosystem provided by the service providing system 5, and receives the provision of the service by accessing the service providing system 5 from the user terminal. Here, the business ecosystem in this embodiment can be a service group composed of a plurality of services that can be logged in with a common user ID, or a service group composed of a plurality of services that can be assigned to users through electronic credits such as common points or the like, or that can be used by users.

The information processing device 1 is a computer having: a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a memory device 14 such as an EEPROM (Electrically Erasable and Programmable Read Only Memory) or an HDD (Hard Disk Drive), and a communication unit 15 such as a NIC (Network Interface Card). However, the specific hardware configuration of the information processing device 1 can be appropriately omitted, replaced, or added according to the implementation. In addition, the information processing device 1 is not limited to a device consisting of a single frame. The information processing device 1 can be implemented by using a plurality of devices such as the so-called cloud or distributed computing technology.

The service providing system 5 is a computer having a CPU, ROM, RAM, a memory device, a communication unit, an input device, an output device, etc. (not shown). Moreover, these systems and terminals are not limited to devices consisting of a single frame. These systems and terminals can be realized by using a plurality of devices such as the so-called cloud or distributed computing technology.

The services provided by the service providing system 5 are, for example, online shopping services, online reservation services, banking services, credit card/postpaid checkout services, electronic money checkout services, advertising services, work center services, or map information services. In addition, "postpaid checkout" is not limited to the so-called Buy Now Pay Later (BNPL) service, and can include any purchase of goods/services due to postpayment. In addition, banking services include: account opening services, deposit management services, financial assistance (asset management) services, and salary transfer services, etc., which can be provided by banks.

The services provided by the service providing system 5 are not limited to the examples in this embodiment. Then, the service providing system 5 notifies the information processing device 1 of the user-related data when providing the service. Here, the user-related data includes the usage record data of the service provided by the user. The content of the service usage record data varies depending on the content of the service, for example, the user's location information record data, credit card usage/postpaid bill usage record data, electronic currency usage record data, transaction record data (including purchase record data of goods, etc.), reservation record data, operation record data from the operation center to the user, etc.

FIG2 is a diagram showing the relationship between the credit score, loyalty score, and comprehensive score in the present embodiment. The information processing device 1 provides the service providing system 5 with the data required to determine whether the target user can be regarded as the target of the provision of the specified service or gift. Previously, the credit score was used as the data required to determine whether the target user can be regarded as the target of the provision of the specified service or gift, and many attempts have been made regarding the determination method or utilization of the credit score. However, in the business ecosystem where a variety of different services are gathered, there is still room for improvement in the user experience of the credit score. Therefore, in this embodiment, the data required to determine whether the target user can be regarded as the target of the provision of the specified service or gift is a credit score and a comprehensive score. Here, the credit score is a score indicating the degree of creditworthiness of the target user; the comprehensive score is a score calculated based on the credit score and the loyalty score. The loyalty score is a score indicating the loyalty of the user described in this embodiment to the business ecosystem, and is calculated based on, for example, the service utilization status of the user.

The service providing system 5 can customize the content of the service or gift provided to the target user in accordance with the data provided from the information processing device 1. For example, the information processing device 1 can provide the service providing system 5 with the credit score or comprehensive score of the target user in accordance with the request from the service providing system 5. The specific form or method of providing data from the information processing device 1 to the service providing system 5 is not limited. For example, the information processing device 1 can provide data to the service providing system 5 in the form of a list of points, etc. In addition, for example, the information processing device 1 can also provide data in the form of returning the score of the target user in accordance with an individual query from the service providing system 5 that specifies the user ID.

Here, the information processing device 1 can make the type of score provided to the service providing system 5 different according to the nature of the service. Specifically, in this embodiment, the information processing device 1 is designed to provide credit scores to the service providing system 5 for services related to credit, and to provide comprehensive scores to the service providing system 5 for services unrelated to credit. Furthermore, in this embodiment, the loyalty score is not provided separately to the service providing system 5. This is because the basic principle of the user score is the credit score, and the loyalty score functions as a bonus element to the credit score within the comprehensive score. However, depending on the implementation form, it can also be designed to provide the loyalty score separately to the service providing system 5.

Furthermore, the information processing device 1 obtains the user attribute data described below from the service providing system 5. The obtained user attribute data is used to generate or update the model required for calculating the credit score, so that the accuracy of the calculated credit score can be improved. Furthermore, by obtaining information on the services used by each user, a part of the comprehensive score, that is, the loyalty score (the score accumulated according to the use of the service by the user) can be updated for each user.

Furthermore, the information processing device 1 notifies the user of the user's comprehensive score. At this time, the user may also be notified of the content of the gift provided to the user by the information processing device 1 or the service providing system 5, and/or the content of the service that will increase the score if the user uses it. In addition, the information processing device 1 does not notify the user of the user's credit score. This is because, in principle, the credit score is only referenced within the system and will not be disclosed to the user.

The information processing device described in the present disclosure uses a machine learning model that takes user attribute data as input and outputs a credit score (a score indicating post-payment risk in the present embodiment). In addition, the attribute data used as input may also include data indicating default (non-performance of debt). By using various attribute data as input, according to the information processing device 1 described in the present disclosure, a universal and generalized credit score generated by uniformly reflecting all attributes of the user can be calculated. The calculated credit score can be used for credit review (determination of approval or rejection of post-payment) in post-payment services, etc., but as mentioned above, in this embodiment, the calculated credit score is also used to determine whether to consider the user as an object of providing a predetermined gift.

Here, the user attribute data includes factual attribute data and inferred attribute data. Attribute data includes, for example, data represented in the form of scores (e.g., continuous values above 0 and below 1) or labels (e.g., binary values corresponding to presence or absence or no). However, the format of attribute data is not limited to the examples in this disclosure. Furthermore, the attribute data may include, for example, online service usage status, and usage status of electronic credits including points. Furthermore, online service usage status may include at least one of the number of cancellations, cancellation rate, and number of orders in online shopping services or online reservation services.

Factual attribute data is data that indicates factual attributes that can be confirmed to be facts about the user based on user-provided data obtained by the user himself or biographical data collected for the user. User-provided data includes, for example, registration data containing the name or email address, phone number, address, workplace, school location, etc. registered by the user himself, or data obtained by the user himself answering a questionnaire. Biographical data includes, for example, the above-mentioned use history data of the e-commerce transaction service provided by the service providing system 5. Factual attribute data is preferably data obtained by converting the above-mentioned user-provided data or biographical data into a data format suitable for marketing and/or analysis purposes. For example, as factual attribute data that can be obtained based on the usage history data, in addition to the types/categories or brands of goods/services that the user frequently uses, it is also possible to cite the commercial districts, entertainment districts, tourist areas, etc. that the user frequently visits.

In addition, the inferred attribute data is data indicating inferred attributes obtained by inference based on user-provided data, resume data, factual attribute data, etc. In this embodiment, the inferred attribute data includes the user's personality, etc., which is inferred or predicted using machine learning technology. The inferred attribute data is preferably data related to attributes that will affect the user's behavior (action) when used in target market selection.

In this embodiment, each attribute data is set with a weight. The weight indicates the degree of correlation between the attribute data and the credit score when the attribute data is used in the calculation of the credit score. Each time the appropriateness of the credit score is evaluated by the machine learning unit 24 described later, the parameters of the model will be adjusted to make the credit score a more appropriate value. The weight corresponding to each attribute data is, for example, equivalent to the weight corresponding to each node (each regression tree) in the model required for the credit score calculation such as the determination model described later, and will be appropriately determined in the process of calculating the credit score. In addition, the credit score, as an example, is determined based on the weight of each node.

Here, the attribute data group may include: demographic attributes, behavioral attributes, or psychographic attributes. Demographic attributes include, for example, the user's gender, family composition, age, etc. Behavioral attributes may be based on the service usage history data, such as: whether electronic cash is used, whether fixed-limit payment is used, deposit and withdrawal history of a specified account, business transaction history of certain goods/services including gambling or lottery (may include online transaction history in online bazaars, etc.), and movement history of users using location information or venue information, etc. Psychographic attributes include, for example, interests in gambling or lottery, etc. However, the attributes of users that can be used are not limited to the examples in this embodiment. For example, "Time required for customer service (outbound calls, etc.)" and "Credit card usage amount/postpaid bill usage amount" from customer service center services can also be used as attributes. Demographic attributes and behavioral attributes can be considered factual attributes. Psychographic attributes can be considered inferred attributes. In addition, attributes similar to demographic attributes may be inferred attributes based on factual attributes based on user-provided data or resume data. Similarly, attributes similar to behavioral attributes may be inferred attributes based on factual attributes based on user-provided data or resume data. Psychographic attributes may be factual attributes obtained based on user-provided data as an example of a result of intention input by the user.

Fig. 3 is a schematic diagram of the functional configuration of the information processing device 1 of the present embodiment. The information processing device 1 functions as an information processing device having a fact attribute determination unit 21, an estimated attribute determination unit 22, a credit score estimation unit 23, a machine learning unit 24, a loyalty score calculation unit 25, a comprehensive score calculation unit 26, a gift determination unit 27, a service identification unit 28, and a notification unit 29 by reading the program recorded in the storage device 14 into the RAM 13 and executing the program by the CPU 11 to control the hardware of the information processing device 1. In addition, in this embodiment and other embodiments described below, the various functions of the information processing device 1 are executed by a general-purpose processor, namely, CPU 11, but some or all of these functions may also be executed by one or more dedicated processors.

The fact attribute determination unit 21 determines fact attribute data about the user that can be confirmed as fact based on the user provided data and/or the resume data of the user provided by the user himself. In this embodiment, the fact attribute determination unit 21 adopts methods such as: summarizing the user provided data and/or resume data, determining the corresponding attributes by referring to other data such as maps, and directly using the user provided data and/or resume data to determine the fact attribute data related to the user. In addition, although the method of determining the fact attribute data related to the user based on the user provided data and/or the resume data of the user is adopted in this embodiment, the fact attribute data related to the user can also be obtained by other methods.

The estimated attribute determination unit 22 determines estimated attribute data estimated for the target user based on at least the user-related data including one or more factual attribute data determined for the target user by the factual attribute determination unit 21. In the present embodiment, the estimated attribute determination unit 22 determines the estimated attribute based on an output value obtained by inputting the user-related data including one or more factual attribute data related to the target user into an attribute estimation model belonging to a machine learning model. In addition, in this embodiment, the output value from the attribute estimation model is a value indicating the probability that the target user has the predetermined estimated attribute. When the output value obtained from the attribute estimation model falls within the predetermined range, the estimated attribute determination unit 22 determines that the target user has the estimated attribute. When it is determined that the target user has the predetermined estimated attribute, the estimated attribute determination unit 22 sets the label of the attribute data estimated for the target user to a value indicating the presence or absence of the attribute or the type of the attribute. In addition, the estimated attribute data may be represented by a score instead of a label. In this case, the estimated attribute determination unit 22 sets a value indicating the degree (probability) to which the estimated attribute can be applied to the score of the attribute data estimated for the target user. The appropriate degree can be the output value of the attribute inference model.

The credit score estimation unit 23 estimates the credit score of the target user (in this embodiment, the credit score calculated based on the post-payment risk) based on the attribute data group including the factual attributes and the estimated attributes related to the target user. Here, the so-called post-payment risk is the risk that the post-payment settlement will not be normally performed by the target user when the target user uses the post-payment settlement, which is represented by a certain index (in this embodiment, the credit score that changes according to the size of the post-payment risk). At this time, the credit score estimation unit 23 may also perform certain processing (normalization, ranking, labeling, etc.) on the determined factual attributes and/or estimated attributes to be used as part of the attribute data group, and may also use other types of scores (such as so-called credit scores, etc.) or labels calculated using the determined factual attributes and/or estimated attributes as all or part of the attribute data group. Here, other machine learning models may also be involved in the calculation of other types of scores or labels.

FIG4 is a simplified diagram of the credit score estimation process described in this embodiment. In this embodiment, the credit score estimation unit 23 estimates (calculates) the credit score of the user by inputting the attribute data group of the user into the credit score estimation model. In this embodiment, the output value of the credit score estimation model is, as an example, a normalized credit score with 400 as the minimum value and 800 as the maximum value.

The machine learning unit 24 generates and/or updates the credit score estimation model used in the credit score estimation performed by the credit score estimation unit 23. The credit score estimation model is a machine learning model that, when one or more attribute data (attribute data group) related to the target user is input, outputs a credit score indicating the degree of risk that the post-payment settlement will not be normally performed by the target user when the target user utilizes the post-payment settlement. By using such a machine learning model, in this embodiment, not only the credit score but also the output taking into account the factors specialized for post-payment can be obtained as the credit score. Furthermore, in this embodiment, an example of a credit score is shown and explained in which the smaller the credit score value, the higher the risk, and the larger the credit score value, the lower the risk, but the relationship between the score value and the level of risk may also be reversed.

When generating and/or updating the credit score estimation model, the machine learning unit 24 creates training data for each user, which defines the attribute data group of the user as an input value and the credit score associated with the user as an output value. Then, the machine learning unit 24 generates and/or updates the credit score estimation model based on the training data. As described above, the attribute data group input to the credit score estimation model includes the factual attribute data determined by the factual attribute determination unit 21 and the estimated attribute data estimated by the estimated attribute determination unit 22 based on the user-related data containing the factual attribute data. The data will be combined with the corresponding user's credit score to become the training data and input to the machine learning unit 24. In this embodiment, the credit score set in the training data is a credit score determined based on the user's payment history data of post-paid billing, which is a combination of user attributes as input values. Here, the payment history data of post-paid billing includes data indicating the presence or absence of default (non-performance of debt) in post-paid billing and the amount of default. At this time, the credit score may be a credit score determined based on a rule, or a credit score that is manually set (annotated). In addition, it may be a credit score that has been output in the past by a credit score estimation model and then modified by an administrator or the like.

The technology involved in the present disclosure can be used as a framework for generating/updating machine learning models used as credit score estimation models, etc. during implementation. As an example, it is based on an ensemble learning algorithm. In the framework, for example, a machine learning framework based on a gradient boosting decision tree (Gradient Boosting Decision Tree: GBDT) (such as LightGBM) can be used. In other words, the framework can adopt a machine learning framework based on such a decision tree model in which the error between the correct answer and the predicted value is inherited between the previous and subsequent weak learners (weak classifiers). The predicted value here, as an example, refers to the predicted value of the credit score. In addition, in addition to LightGBM, the framework can also adopt boosting techniques such as XGBoost or CatBoost. If a framework using a decision tree is used, a machine learning model with relatively high performance can be generated/updated with fewer parameter adjustments than a framework using a neural network. However, the machine learning model generation/update framework that can be used when implementing the technology described in this disclosure is not limited to the examples in this embodiment. For example, other learners such as random forests can be used instead of gradient boosting decision trees as learners, and learners such as neural networks that are not called so-called weak learners can also be used. In addition, especially when a learner such as a neural network that is not called a so-called weak learner is used, ensemble learning may not be used.

FIG5 is a simplified diagram of the concept of a decision tree of a machine learning model used as a credit score estimation model in the present embodiment. In the case of using a machine learning framework of gradient boosting based on a decision tree algorithm, optimization of divergence conditions of each node of the decision tree is performed. Specifically, in a machine learning framework of gradient boosting based on a decision tree algorithm, credit scores are calculated for user groups having attributes represented by each of two child nodes diverged from a parent node, and the divergence conditions of the parent node are optimized so that the difference in the credit score becomes larger (for example, the difference becomes the maximum, or becomes above a predetermined threshold), that is, the two child nodes can diverge clearly. For example, if the attribute represented as the divergence condition of the node is age, the age set as the divergence threshold can be changed, or the divergence condition can be changed to an attribute other than age. In this way, by recursively optimizing the divergence conditions of all nodes in the decision tree, the accuracy of credit score estimation based on the attribute data group can be improved.

The loyalty score calculation unit 25 calculates the user's loyalty score with respect to the business ecosystem based on the service utilization status of the user in the business ecosystem described in this embodiment. In this embodiment, the loyalty score is a score with 0 as the minimum value and 200 as the maximum value. The higher the user's loyalty to the target business ecosystem, the larger the value will be output. Here, the service utilization status is not limited to whether the user has utilized the service, but may also include business indicators such as the transaction amount and/or login frequency of each service performed by the user. The loyalty score calculation unit 25 can, for example, calculate the final loyalty score based on a predetermined threshold and in accordance with the business indicators of each service.

Specifically, in this embodiment, the loyalty score calculation unit 25 calculates the loyalty score of the user using a calculation formula that is pre-set for each service that the user can use in the business ecosystem. For example, the loyalty score calculation unit 25 calculates the loyalty point of the target user by adding up the points set for the services whose usage history of the target user is recorded during the set target period. More specifically, "add 10 points" is set for service (1), "add 20 points" is set for service (2), and "add 30 points" is set for service (3). If the target user has used service (1) and service (3) during the target period and this fact is recorded in the user database, the loyalty score calculation unit 25 adds the 10 points set for service (1) and the 30 points set for service (3) recorded in the user's usage history, and sets the loyalty score of the user to 40.

Furthermore, the calculation method of the loyalty score is not limited to the above-mentioned example. For example, points calculated based on the current score (credit score, loyalty score or comprehensive score) may be added. Specifically, assuming that "1% of the comprehensive score" is set for service (4), and the fact that the user has used service (4) is recorded in the user database, the loyalty score calculation unit 25 multiplies the user's current comprehensive score by 1% and adds the calculated value to the loyalty score.

Furthermore, for example, the value or ratio of the points added may also be different for each user. Specifically, assuming that "10 points are added for the first-time user and 1 point is added for the second and subsequent users" for service (5), when the user uses service (5) for the first time, the loyalty point calculation unit 25 adds 10 points to the user's loyalty point, and when the user uses service (5) again, the loyalty point calculation unit 25 adds 1 point to the user's loyalty point. In addition, the method of calculating the loyalty point can adopt various methods, and is not limited to the above example.

The comprehensive score calculation unit 26 calculates the comprehensive score of the target user based on the credit score estimated for the target user and the loyalty score calculated for the target user. At this time, the comprehensive score calculation unit 26 calculates the comprehensive score of the target user in such a way that the weight (ratio) of the loyalty score in the comprehensive score is smaller than the weight (ratio) of the credit score in the comprehensive score. In this embodiment, the comprehensive score is calculated by simply adding the credit score and the loyalty score for the target user. Therefore, in this embodiment, the upper limit of the loyalty score (e.g., 200) is set to a value that does not exceed the lower limit of the credit score (400 in this embodiment) by a predetermined ratio, so that the weight of the loyalty score in the comprehensive score is less than the weight of the credit score in the comprehensive score, thereby calculating the comprehensive score of the target user. However, the specific method required for calculating the comprehensive score is not limited to the example in this embodiment. For example, the credit score and loyalty score may be multiplied by their respective weights C and L to calculate a comprehensive score (comprehensive score = credit score * weight C + loyalty score * weight L), and the weights C and L may be set so that the proportion of the loyalty score in the comprehensive score is smaller than the proportion of the credit score in the comprehensive score.

The gift determination unit 27 determines the content of the gift to the target user based on the target user's comprehensive score or the target user's credit score. In this embodiment, the score used in the determination is differentiated according to the type of the gift. That is, in this embodiment, the gift determination unit 27 determines the content of the gift to the target user related to the credit based on the target user's credit score, and determines the content of the gift to the target user not related to the credit based on the target user's comprehensive score. With this design, for gifts that need to be granted based solely on the user's creditworthiness (such as permission to pay for expenses later and interest rate discounts, gifts related to credit), the decision on whether to grant the gift is based on the credit score; for gifts that can reflect the user's intention (it is okay to use a score that can be intentionally changed by the user) (such as waived bank fees, gifts not related to credit), the decision on whether to grant the gift can be based on the comprehensive score.

Specifically, the gift determination unit 27 compares the threshold value set in advance for each gift with the current score of the target user, and when the score is above the threshold value, it is determined that the target user can be provided with the gift related to the threshold value. For example, when the threshold value set for the interest rate discount gift is "credit score 700", the interest rate discount gift will be provided to users with a credit score of 700 or above. Here, since the score referred to is the credit score, even if the user intentionally uses the service of the target score improvement, this alone cannot improve the credit score, so the user must use other methods to improve the credit score. For another example, if the threshold for a gift of free bank fees is set to "comprehensive score of 700", a gift of free bank fees will be provided to users with a comprehensive score of 700 or above. Here, since the score being referenced is the comprehensive score, the user can intentionally use the score to improve the target service, and the user can intentionally operate the comprehensive score to receive the gift.

The service identification unit 28 identifies services that can improve the target user's loyalty score and overall score when the target user uses them, from among the services that can be used by the user in the business ecosystem. Specifically, the service identification unit 28 extracts services that have been set up so that the user satisfies the conditions for use and will be given a loyalty score by using the services from a list of services that is pre-stored in a database. Here, the specific content of the conditions for use is not limited, but for example, user attributes such as the user's age, place of residence, and service use registration status can be referenced as conditions for use. Furthermore, the conditions for use may also include conditions other than user attributes (such as the service provision period). In addition, since the credit score and loyalty score are calculated in different ways, the value added to the loyalty score set for each service can be set arbitrarily within the range of the pre-set value by the service provider or system administrator. For example, for a service that you particularly want users to use (and that you are promoting hard), you can set the loyalty score to a larger value to increase user demand for the service.

The notification unit 29 notifies the target user of the comprehensive score of the target user, the content of the gift determined for the target user, and the service specified by the service specifying unit 28.

FIG6 is an example of the first notification screen displayed on the user terminal of the target user in this embodiment. The first notification screen is displayed through the application or browser used by the user who logs into the business ecosystem, and contains the score of the target user (765 points in the example shown in the figure) and a list of services for which a discount (gift) will be received (banking services and electronic money checkout services in the example shown in the figure). In addition, the score notified to the user is only the comprehensive score, and its details (i.e., credit score and loyalty score) are not notified. For example, the details of the score of 765 points shown in the figure are, for example, the total of a credit score of 750 points and a loyalty score of 15 points, but the user will not be informed of the details. Moreover, in the first notification screen, the user can select the service display portion in the overview by touching or clicking, so that the second notification screen required for browsing the gifts and score-enhancing services that will be received in the selected service can be displayed.

FIG. 7 is an example of a second notification screen displayed on the user terminal of the target user in the present embodiment. The second notification screen is also displayed by the application or browser used by the user who logs into the business ecosystem, similarly to the first notification screen. However, the application used for display here may be an application different from the application displayed on the first notification screen (an application corresponding to the discount content). Specifically, FIG. 7 illustrates a second notification screen displayed by the banking service application when the display portion of the banking service is selected in the first notification screen of FIG. 6 and the application is migrated to the banking service application.

The second notification screen includes: the target user's score, the gift provided to the target user (displayed as the item "Promotional content" in the figure), and the service that can improve the target user's score (displayed as the item "Want to improve the score?" in the figure). In the example shown in the figure, as the item "Promotional content", it is displayed that the target user will be provided with interest rate discounts on banking services and free banking service fees. Then, as the item "Want to improve the score?", it is suggested to use the financial assistance (asset management) service (displayed as "Financial support" in the figure) or the lottery purchase service. By using the proposed service, the target user's points will increase, and the gift provided to the target user will have the possibility of increasing. Therefore, such a display will become a promotion for users who want to increase gifts. In addition, in the notification screen, the user can select the gift (discount content) or the service display part required for point increase by touching or clicking, so that the detailed introduction screen of the gift or service or the screen required for using the gift or service is displayed (illustration omitted).

In addition, in this embodiment, the example in which the notification unit 29 directly notifies the user of the value of the comprehensive score is given for explanation, but the notification unit 29 may also replace the value of the comprehensive score with an indicator based on the comprehensive score of the target user to notify the target user. The indicator used here may be, for example, a user class set for each range of the score value. The expression of the user class may also be appropriately selected according to the embodiment, and may be, for example, the number of stars, a value, a text, a word (such as copper, silver, gold, etc.), etc.

<Processing flow> Next, the process flow of the processing performed by the information processing device described in this embodiment is described. In addition, the specific content and processing order of the processing described below are an example required for implementing the present disclosure. The specific processing content and processing order can be appropriately selected according to the implementation form of the present disclosure.

Fig. 8 is a flow chart showing the flow of the machine learning process according to the present embodiment. The process shown in this flow chart is executed periodically or at a time point specified by an administrator.

In this embodiment, a credit score estimation model is generated and/or updated during the machine learning process. The machine learning unit 24 creates training data, which includes: a group of attribute data of each user accumulated in the past, and a combination of credit scores predetermined for the corresponding user (step S101). Then, the machine learning unit 24 inputs the created training data into the credit score estimation model, and generates and/or updates the credit score estimation model used when the credit score estimation unit 23 estimates the credit score (step S102). Thereafter, the process shown in this flowchart ends.

Fig. 9 is a flow chart showing the flow of the score calculation process according to the present embodiment. The process shown in this flow chart is executed for each target user periodically or at a designated time point.

In step S201 and step S202, factual attribute data and inferred attribute data are determined. The factual attribute determination unit 21 determines the factual attribute data related to the target user based on the user-provided data and/or resume data of the target user (step S201). Then, the inferred attribute determination unit 22 determines the inferred attribute data related to the target user based on at least the factual attribute data determined in step S201 (step S202). Thereafter, the process proceeds to step S203.

In step S203 and step S204, the credit score is determined and output. The credit score estimation unit 23 determines an attribute data group including the factual attribute data determined in step S201 and the estimated attribute data determined in step S202 (step S203). Then, the credit score estimation unit 23 inputs the attribute data group determined in step S203 into the credit score estimation model, and obtains the output value as a credit score that changes in accordance with the risk that the settlement of the post-payment settlement is not normally performed by the target user when the target user uses the post-payment settlement (step S204). However, the credit score estimation method is not limited to the example in this embodiment. For example, the credit score may also be a value calculated by inputting the attribute data group into a predetermined function or statistical model other than the machine learning model. Thereafter, the process proceeds to step S205.

In step S205 and step S206, the loyalty score and the comprehensive score are calculated. The loyalty score calculation unit 25 calculates the loyalty score of the user with respect to the business ecosystem based on the service utilization status of the user in the business ecosystem (step S205). Then, the comprehensive score calculation unit 26 calculates the comprehensive score of the target user based on the credit score estimated in step S204 for the target user and the loyalty score calculated in step S205 for the target user (step S206). In this embodiment, for the target user, the credit score and the loyalty score are simply added together to calculate a comprehensive score. Then, the process proceeds to step S207.

In step S207 and step S208, the gift provided to the target user is determined. The gift determination unit 27 determines the content of the gift related to credit that can be provided to the target user based on the credit score of the target user (step S207). Furthermore, the gift determination unit 27 determines the content of the gift not related to credit that can be provided to the target user based on the comprehensive score of the target user (step S208). Thereafter, the process proceeds to step S209.

In step S209, services that contribute to the improvement of the target user's score are identified. The service identification unit 28 identifies services that can be used by the target user and that can improve the target user's loyalty score and comprehensive score if the target user uses them. Thereafter, the process proceeds to step S210.

In step S210, the target user is notified. The notification unit 29 sends the target user's comprehensive score calculated in step S206, the gift content determined for the target user in steps S207 and S208, and the service specified for the user in step S209 to the target user's terminal, and displays the information on the user's terminal to notify the target user. Thereafter, the processing shown in this flowchart ends.

<Effect> According to this implementation form, in the business ecosystem (economic circle) where various types of services are gathered, the user evaluation method and user experience can be improved.

<Variation> In the above-described implementation form, the service that contributes to the improvement of the user's score is specified and notified to the target user. However, in order to make a stronger appeal to the user to use the service, it is also possible to design a service that can add gift content to the target user when the user uses it, and notify the user. In addition, in this variation, the configuration common to the above-described implementation form is omitted, and the differences are explained.

Fig. 10 is a diagram schematically showing the functional configuration of the information processing device 1b according to the present modification. The information processing device 1b functions as an information processing device having a loyalty score calculation unit 25, a comprehensive score calculation unit 26, a gift determination unit 27, a service identification unit 28b, a notification unit 29, and a score update prediction unit 30 by reading the program recorded in the storage device 14 into the RAM 13 and executing the program by the CPU 11 to control the hardware of the information processing device 1. That is, the information processing device 1b described in the embodiment further includes a score update prediction unit 30 in addition to the configuration of the information processing device 1, and the processing content of the service identification unit 28b is partially different from that of the above embodiment.

The score update prediction unit 30 calculates the target user's loyalty score and comprehensive score (hereinafter referred to as "estimated comprehensive score") for each service that the user can use in the business ecosystem, assuming that the target user has used the service. The calculation method of the estimated comprehensive score is roughly the same as the calculation method of the comprehensive score in the above-described embodiment, but the score update prediction unit 30 calculates the estimated comprehensive score of the target user assuming that the user has used the service for each service that the user has not used.

The service specifying unit 28b specifies the service that can be added as a gift content to the target user if the target user uses it, based on the prediction result made by the score update prediction unit 30. The service specifying unit 28b compares the threshold value preset for each gift with the estimated comprehensive score of the target user for the gifts that have not yet been provided to the user. Then, when the estimated comprehensive score is above the threshold value, the service specifying unit 28b specifies the service related to the estimated comprehensive score as a service that can be added as a gift content to the target user if the target user uses it. For example, if the threshold for a gift of waived bank fees is set to "estimated overall score of 700", for users whose current overall score is less than 700 but whose estimated overall score will be above 700 if they use a certain service, the service will be identified as a service that will provide the user with a gift if the user uses it.

The notification unit 29 notifies the target user of the service identified by the service identification unit 28b as described above. By designing in this way, the user can know the specific service that can be used to obtain a gift, and the user can be more effectively urged to use the service.

In addition, in the implementation form described above, a score indicating post-payment risk is used as an example for explanation as a credit score, but a credit score can be any score that indicates the creditworthiness of a user, and can be a score that indicates the creditworthiness of a user based on other indicators that are not post-payment risk (such as credit card default risk or loan default risk, etc.).

1,1b: Information processing device 5: Service provision system 11: CPU 12: ROM 13: RAM 14: Memory device 15: Communication unit 21: Factual attribute determination unit 22: Estimated attribute determination unit 23: Credit score estimation unit 24: Machine learning unit 25: Loyalty score calculation unit 26: Comprehensive score calculation unit 27: Gift determination unit 28,28b: Service identification unit 29: Notification unit 30: Score update prediction unit

[Figure 1] A schematic diagram of the configuration of the information processing system according to the embodiment. [Figure 2] An illustration of the relationship between the credit score, loyalty score, and comprehensive score in the embodiment. [Figure 3] An illustration of the schematic functional configuration of the information processing device according to the embodiment. [Figure 4] A simplified diagram of the credit score estimation process according to the embodiment. [Figure 5] A simplified diagram of the concept of the decision tree of the machine learning model adopted as the credit score estimation model, etc. in the embodiment. [Figure 6] An example of the first notification screen displayed on the user terminal of the target user in the embodiment. [Figure 7] An example of the second notification screen displayed on the user terminal of the target user in the embodiment. [Figure 8] A flow chart of the flow of the machine learning process according to the embodiment. [Figure 9] A flow chart of the score calculation process described in the embodiment. [Figure 10] A schematic diagram of the functional structure of the information processing device described in the variant.

1: Information processing device

21: Determination of the nature of facts

22: Presumed attribute determination section

23: Credit score estimation department

24: Machine Learning Department

25: Loyalty score calculation department

26: Comprehensive score calculation unit

27: Gift Decision Department

28: Service specific department

29: Notification Department

Claims (16)

An information processing system comprises: Credit score estimation means for estimating a credit score to be set for a user based on a group of attribute data related to the user; and Loyalty score calculation means for calculating the user's loyalty score with respect to a predetermined business ecosystem based on the service utilization status of previous users in the predetermined business ecosystem; and Comprehensive score calculation means for calculating the comprehensive score of the target user based on the previous credit score estimated for the target user and the previous loyalty score calculated for the target user. The information processing system as described in claim 1, wherein, further comprises: gift determination means for determining the content of the gift to be given to the target user based on the comprehensive score of the target user or the credit score of the target user. The information processing system as described in claim 2, wherein the pre-credit gift determination means determines the content of the gift to the pre-credit target user related to the credit based on the credit score of the pre-credit target user, and determines the content of the gift to the target user not related to the credit based on the comprehensive score of the target user. The information processing system as described in claim 2, wherein, it also has: a notification means for notifying the target user of the content of the pre-recorded gift determined for the target user of the pre-recorded gift. The information processing system as described in claim 1, wherein, further comprises: a notification means for notifying the target user of the previous record comprehensive score of the target user or an indicator based on the comprehensive score. The information processing system as described in claim 1, further comprising: Service identification means for identifying services that can be used by the foreword users in the foreword business ecosystem, which, when used by the foreword target users, can improve the foreword loyalty score and the foreword comprehensive score of the target users; and Notification means for notifying the foreword target users of the foreword services identified by the foreword service identification means. The information processing system as described in claim 6, wherein, further comprises: score update prediction means, which is used to calculate the previous record comprehensive score of the target user assuming that the target user of the previous record has used the service according to each service that the target user of the previous record can use in the previous record business ecosystem; previous record service specific means, which is based on the prediction result made by the previous record score update prediction means, and specifies the service that can add the previous record gift content to the target user if the target user of the previous record uses the service. The information processing system as described in claim 1, wherein the aforementioned attribute data group includes: attribute data collected due to the aforementioned user utilizing the aforementioned service in the aforementioned business ecosystem. The information processing system as described in claim 1, wherein the pre-record comprehensive score calculation means is to calculate the comprehensive score of the pre-record target user in such a way that the weight of the pre-record loyalty score in the pre-record comprehensive score is less than the weight of the pre-record credit score in the pre-record comprehensive score. An information processing system as recited in claim 1, wherein the pre-credit score estimation means uses: a credit score estimation model generated and/or updated using training data that uses a pre-credit attribute data group as an input value and a pre-credit score determined based on payment records of post-payment related to a common user of the attribute data group as an output value, to estimate the credit score to be set for the pre-credit user. An information processing system as described in claim 10, wherein the output value of the pre-credit score estimation model is a score that changes according to the post-payment risk that the post-payment settlement is not normally performed by the user when the pre-credit user uses the post-payment settlement. An information processing system as described in claim 10, wherein the pre-credit score estimation means uses: a pre-credit score estimation model generated and/or updated using a machine learning framework based on a gradient boosting decision tree to estimate the pre-credit score. The information processing system as described in claim 1, wherein, further comprises: fact attribute determination means for determining fact attributes that can be confirmed as facts for the user based on user-provided data provided by the previous user himself or the resume data of the user; and inferred attribute determination means for determining inferred attributes inferred for the user based on at least the previous fact attributes related to the user; previous credit score inference means for inferring the credit score of the previous user based on an attribute data group containing the previous fact attributes and inferred attributes related to the previous user. The information processing system as described in claim 1, wherein the aforementioned loyalty score calculation means uses a calculation formula pre-set according to each service that the aforementioned user can use in the aforementioned business ecosystem to calculate the user's loyalty score. An information processing method is performed by a computer: A credit score estimation step is to estimate the credit score to be set to the user based on the attribute data group related to the user; and A loyalty score calculation step is to calculate the user's loyalty score with respect to the business ecosystem based on the service utilization status of the previous user in the specified business ecosystem; and A comprehensive score calculation step is to calculate the comprehensive score of the target user based on the previous credit score estimated for the target user and the previous loyalty score calculated for the target user. A program product is a computer that functions to: Credit score estimation means for estimating a credit score to be set for a user based on a group of attribute data associated with the user; and Loyalty score calculation means for calculating the user's loyalty score with respect to a predetermined business ecosystem based on the service utilization status of previous users in the business ecosystem; and Comprehensive score calculation means for calculating a comprehensive score for a target user based on a previous credit score estimated for the target user and a previous loyalty score calculated for the target user.

Images