Estimation system, estimation method, and program

Abstract

Streamline actions for the target subjects. 【Solution】The effect information acquisition unit (102) of the estimation system (1) acquires effect information regarding the effects obtained when a predetermined action is performed on each of a plurality of target subjects by a predetermined means at a predetermined timing. The cost information acquisition unit (103) acquires cost information regarding the costs required when an action is performed on each of the plurality of target subjects by a means at a timing. The constraint information acquisition unit (104) acquires constraint information regarding the constraints on the costs allowed for the action. The estimation unit (105) estimates, based on the effect information and cost information of each of the plurality of target subjects and the constraint information, a combination of the target subject on whom the action is to be performed, and at least one of the timing and the means, such that the effects can be efficiently obtained as a whole within the range of the constraints.

Description

Technical Field

[0001] The present disclosure relates to an estimation system, an estimation method, and a program.

Background Art

[0002] Conventionally, actions such as reminders have been taken against those who have overdue credit card payments. For example, in Patent Document 1, based on a learning model prepared using machine learning techniques, among a plurality of subjects who have overdue credit card payments, an operator of a card company estimates subjects for whom reminders such as phone calls are highly effective, and a system in which the operator makes reminders to the subjects is described.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the technology of Patent Document 1 can estimate the effectiveness of reminders for each of a plurality of subjects who have overdue credit card payments, but cannot perform an estimation focusing on the overall efficiency of reminders. This is the same in other scenarios different from the reminder of credit card payments such as Patent Document 1. Conventional technologies have not been able to determine the necessity of actions against subjects in consideration of overall efficiency.

[0005] One of the objects of the present disclosure is to improve the efficiency of actions against subjects.

Means for Solving the Problems

[0006] The estimation system according to the present disclosure includes an effect information acquisition unit that acquires effect information regarding an effect obtained when a predetermined action is performed on each of a plurality of subjects by a predetermined means at a predetermined timing, a cost information acquisition unit that acquires cost information regarding a cost required when the action is performed on each of the plurality of subjects by the means at the timing, a constraint information acquisition unit that acquires constraint information regarding a constraint on a cost allowed for the action, and an estimation unit that estimates a combination of the subject on which the action is to be performed, and at least one of the timing and the means, so that the effect can be efficiently obtained as a whole within the range of the constraint, based on the effect information and the cost information of each of the plurality of subjects, and the constraint information.

Advantages of the Invention

[0007] The present disclosure can improve the efficiency of actions on subjects.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0009] [1. Hardware Configuration of the Deduction System] An example of an embodiment of a deduction system, a deduction method, and a program according to the present disclosure will be described. FIG. 1 is a diagram showing an example of the hardware configuration of the deduction system. For example, the deduction system 1 includes a server 10, an operator terminal 20, and a target terminal 30. Each of the server 10, the operator terminal 20, and the target terminal 30 is connected to a network N such as the Internet, a LAN, or a public telephone line. In FIG. 1, each of the server 10, the operator terminal 20, and the target terminal 30 is shown as one unit, but at least one of the server 10, the operator terminal 20, and the target terminal 30 may exist in a plurality of units.

[0010] The server 10 is a server computer. For example, the server 10 includes a control unit 11, a storage unit 12, and a communication unit 13. The control unit 11 includes at least one processor. The storage unit 12 includes at least one of a volatile memory such as a RAM and a non-volatile memory such as a flash memory. The communication unit 13 includes at least one of a communication interface for wired communication and a communication interface for wireless communication.

[0011] The operator terminal 20 is the terminal of the operator described later. For example, the operator terminal 20 is a smartphone, a tablet, a mobile phone not classified as a smartphone, a landline phone, a personal computer, or a wearable terminal. The operator terminal 20 includes a control unit 21, a storage unit 22, a communication unit 23, an operation unit 24, and a display unit 25. The hardware configurations of the control unit 21, the storage unit 22, and the communication unit 23 may be the same as those of the control unit 11, the storage unit 12, and the communication unit 13, respectively. The operation unit 24 is an input device such as a touch panel or a mouse. The display unit 25 is a display such as a liquid crystal or an organic EL.

[0012] The target terminal 30 is the terminal of the target person described later. For example, the target terminal 30 is a smartphone, a tablet, a mobile phone not classified as a smartphone, a landline phone, a personal computer, or a wearable terminal. For example, the target terminal 30 includes a control unit 31, a storage unit 32, a communication unit 33, an operation unit 34, and a display unit 35. The hardware configurations of the control unit 31, the storage unit 32, the communication unit 33, the operation unit 34, and the display unit 35 may be the same as those of the control unit 11, the storage unit 12, the communication unit 13, the operation unit 24, and the display unit 25, respectively.

[0013] Note that the programs stored in the storage units 12, 22, and 32 may be supplied to the server 10, the person-in-charge terminal 20, or the target terminal 30 via the network N. Further, at least one of a reading unit (for example, a memory card slot) that reads a computer-readable information storage medium and an input / output unit (for example, a USB port) for inputting / outputting data with an external device may be included in the server 10, the person-in-charge terminal 20, or the target terminal 30. For example, a program stored in the information storage medium may be supplied to the server 10, the person-in-charge terminal 20, or the target terminal 30 via at least one of the reading unit and the input / output unit.

[0014] Also, the estimation system 1 may include at least one computer. The computer included in the estimation system 1 is not limited to the example of FIG. 1. For example, the estimation system 1 may include only the server 10 and the person-in-charge terminal 20. In this case, the target terminal 30 exists outside the estimation system 1. The estimation system 1 may include only the server 10. In this case, the person-in-charge terminal 20 and the target terminal 30 exist outside the estimation system 1. For example, the estimation system 1 may include the server 10 and another computer not shown in FIG. 1.

[0015] [2. Outline of Estimation System] In this embodiment, for each of a plurality of target persons, a predetermined action is performed by a predetermined means at a predetermined timing. A target person is a person who is the target of a predetermined action. In this embodiment, since a predetermined action may not be performed on a target person, a target person can also be said to be a candidate person on whom a predetermined action is to be performed. For example, a target person is a person who uses a predetermined service. The predetermined service is a service in which there is a possibility that some action on the target person may occur.

[0016] In this embodiment, a case where a credit card service corresponds to the predetermined service is taken as an example. Therefore, the parts that describe the credit card service can be read as the predetermined service. The predetermined service may be any service. The predetermined service is not limited to the credit card service. For example, the predetermined service may be an e-commerce service, a payment service, a travel reservation service, a communication service, a financial service, an online free market service, or other services. A target person does not have to be using the predetermined service. A target person may be a potential customer in the predetermined service.

[0017] The predetermined timing is the timing at which a predetermined action is performed. The predetermined timing may be a pinpoint moment, or may have a certain length (for example, 1 hour, 1 day, or 1 week). In this embodiment, a case where the predetermined timing is indicated by a date is taken as an example, but the predetermined timing may be indicated by a date and time including not only the date but also the time. The predetermined timing may be indicated only by the time without including date information. The predetermined timing may be indicated by other information other than the date and time. The predetermined timing may be indicated by temporal information. For example, the predetermined timing may be indicated by a day of the week, weekdays, weekends, the beginning of the month, the end of the month, the first ten days, the middle ten days, the last ten days, or other information. There may be only one predetermined timing, or there may be a plurality of predetermined timings.

[0018] The predetermined means is the means used for a predetermined action. In other words, the predetermined means is the specific method of the predetermined action. There may be only one predetermined means, or there may be a plurality of predetermined means. For example, when contacting a target person corresponds to a predetermined action, the predetermined means is a contacting means. The contacting means may be various known means. For example, the contacting means may be human-to-human calling, robot-to-human calling by a robot, SMS (Short Message Service), email, SNS (Social Networking Service), message apps, chat, notification functions of applications such as smartphone apps, or other means.

[0019] Note that the predetermined means is not limited to the contacting means. The predetermined means may be any means corresponding to the predetermined action. For example, when presenting an advertisement corresponds to a predetermined action, the predetermined means is an advertising medium used as an advertisement. For example, the advertising medium may be a banner, push, pop-up, email, paper direct mail, public broadcasting, digital signage, or other media. When the operation of a predetermined service corresponds to a predetermined action, the predetermined means is a business method used in the business. For example, the business method may be the various contacting means described above, direct visits, setting up a booth at an event venue, or other means. When the predetermined action is an action other than contacting, presenting an advertisement, and business, the predetermined means may be any means used in the other action.

[0020] The predetermined action is an act performed on the target person. The predetermined action may be an act performed for any purpose. In this embodiment, a case where contact with the target person corresponds to the predetermined action is taken as an example. The predetermined action may be any act. The predetermined action is not limited to contact with the target person. For example, the predetermined action may be the presentation of an advertisement or sales activities. The predetermined action may be other acts other than contact, advertisement presentation, and sales activities. The person who performs the predetermined action may be any person. The person who performs the predetermined action may be a person other than the target person. For example, the person who performs the predetermined action may be an employee of an operator of a predetermined service, an employee of another operator who cooperates with the operator, a part-time worker, or others.

[0021] In this embodiment, a case where the estimation system 1 is used in the scene of debt collection in a credit card service is taken as an example. For example, each of a plurality of target persons is a delinquent who has defaulted on credit card payments. A delinquent is a person who could not have a credit card debited by the due date due to insufficient balance in a bank account or the like. The predetermined means is a means of contact for urging the delinquent. The part that explains the means of contact in this embodiment can be read as the predetermined means. Urging is a reminder of the payment defaulted by the target person. Since the predetermined means is a means used in urging, it can also be called an urging means. Since the predetermined means is a means used in debt collection, it can also be called a collection means. The predetermined action is an urging of the delinquent. The part that explains the urging in this embodiment can be read as the predetermined action.

[0022] In this embodiment, an example is given where a person in charge working for the credit card company that issued the credit card makes a demand on the target person. Further, an example is given where the person in charge can use three communication means: personal call, robot call, and SMS. The target person may or may not respond to the communication from the person in charge. Just because the target person responds to the communication from the person in charge does not mean that the overdue payment will be made. Conversely, even if the target person does not respond to the communication from the person in charge, the overdue payment may still be made.

[0023] For example, the administrator who manages the person in charge plans how to make a demand on which target person, at which timing, and by which communication means so as to recover more debts. On the other hand, for the person in charge to make a demand on the target person, it costs according to the communication means. In this embodiment, the cost is the cost for making a demand. The cost is a burden on the credit card company. For example, the cost may be a monetary cost, a time cost, the number of demands, the cost of hardware resources, the cost of software resources, or other costs.

[0024] For example, when a demand is made by a personal call, the cost includes labor costs and telephone charges. When a demand is made by a robot call, the cost includes telephone charges. When the credit card company uses a paid robot call service, the cost includes the usage fee of the robot call service. When a demand is made by SMS, the cost includes the SMS usage fee. The cost allowed by the credit card company varies depending on the budget, human resources, hardware resources, software resources, or other circumstances. The person in charge needs to efficiently recover debts within the scope of cost constraints.

[0025] However, even if the administrator tries to plan which target person to urge, at what timing, and by which means of communication, there are innumerable such combinations. Hereinafter, these combinations are referred to as urging combinations. Since there are innumerable urging combinations, it is very difficult for the administrator to make an efficient plan. Even if the collection of claims from a specific target person is successful, if the collection of claims is not streamlined for the entire credit card service, the card company will incur losses because it has to cover the delinquency of the target person.

[0026] Therefore, the estimation system 1 of the present embodiment estimates an urging combination so that the collection of claims is streamlined for the entire credit card service within the range of cost constraints allowed for the card company. The person in charge urges the target person based on the urging combination estimated by the estimation system 1. As a result, the estimation system 1 can streamline the urging of the target person. Hereinafter, the details of the estimation system 1 will be described.

[0027] [Functions Realized by the Estimation System] FIG. 2 is a diagram showing an example of functions realized by the estimation system 1. In the present embodiment, the case where the main functions of the estimation system 1 are realized by the server 10 is taken as an example. For example, the server 10 includes a data storage unit 100, a probability estimation model storage unit 101, an effect information acquisition unit 102, a cost information acquisition unit 103, a constraint information acquisition unit 104, and an estimation unit 105. Each of the data storage unit 100 and the probability estimation model storage unit 101 is realized by the storage unit 12. Each of the effect information acquisition unit 102, the cost information acquisition unit 103, the constraint information acquisition unit 104, and the estimation unit 105 is realized by the control unit 11.

[0028] [Data Storage Unit] The data storage unit 100 stores data necessary for estimating the urging combination. For example, the data storage unit 100 stores a target person database DB.

[0029] FIG. 3 is a diagram showing an example of a target person database DB. The target person database DB is a database in which various data regarding target persons are stored. For example, the target person database DB stores a target person ID, a credit card number, personal information, demographic information, delinquency information, and usage status information. Note that any data may be stored in the target person database DB. The data stored in the target person database DB is not limited to the example of FIG. 3. For example, the target person database DB may store data of users who have not made a late payment (i.e., users who do not correspond to the target persons).

[0030] The target person ID is an example of target person identification information that can identify a target person. Therefore, the places where the target person ID is described can be read as target person identification information. The target person identification information may be other information than the target person ID. The target person identification information may be any information that can identify the target person in some form. For example, the target person identification information may be the target person's email address, phone number, or other information. The credit card number may be used as the target person identification information. Note that the target person ID may be used as a login account for the credit card service.

[0031] The personal information stored in the target person database DB is the personal information of the target person. For example, the personal information indicates the target person's name, address, phone number, email address, or a combination thereof. The demographic information is information regarding the characteristics of the target person. The personal information may correspond to the demographic information, or the demographic information may correspond to the personal information. For example, the demographic information indicates the target person's gender, age or age group, occupation, annual income, family composition, or a combination thereof. Each of the personal information and the demographic information may be publicly known information.

[0032] The delinquency information is information regarding the payments delinquent by the target person. For example, the delinquency information is the delinquency amount, the breakdown of the payments delinquent by the target person, the date of debit of the payments delinquent by the target person, or a combination thereof. The delinquency amount is the total amount of the payments delinquent by the target person. The breakdown of the payments delinquent by the target person is the details of the settlement indicated by the said payments. For example, the breakdown of the payments delinquent by the target person is the settlement date, settlement amount, settlement location (e.g., store, etc.) of the credit card used by the target person, or a combination thereof. When a certain target person becomes delinquent in a payment, the server 10 updates the delinquency information associated with the target person ID of the said target person.

[0033] The usage information is information regarding the usage status of the credit card by the target person. The usage information may also include information regarding payments not delinquent by the target person. The usage information may be the usage details of the credit card. For example, the usage information indicates the settlement date, settlement amount, settlement location (e.g., store, etc.) of the credit card used by the target person, or a combination thereof. The usage information may indicate installment payments, bonus payments, or other usage statuses of the credit card. The usage information may indicate the payment amount of the target person for each month. When a target person uses the credit card service, the server 10 updates the usage information associated with the target person ID of the said target person.

[0034] Note that the data stored in the data storage unit 100 is not limited to the target person database DB. The data storage unit 100 may store the data necessary for the estimation of the combination of reminders. For example, the data storage unit 100 may store a training database necessary for the learning of the probability estimation model M1 described later. The data storage unit 100 may store a management tool for the persons in charge of the credit card company.

[0035] [Probability Estimation Model Storage Unit] The probability estimation model storage unit 101 stores a probability estimation model M1 in which the relationship between the subject characteristic information regarding the characteristics of the subject for training and the probability of obtaining an effect from the subject for training is learned. The subject for training may be an actual subject or a fictional subject. The subject for training may be a subject of another credit card service different from the credit card service for which debt collection is planned.

[0036] The subject characteristic information is information regarding the characteristics of the subject. The subject characteristic information indicates characteristics that have a correlation with the probability of obtaining an effect when a reminder is issued. The correlation is learned in the probability estimation model M1. The subject characteristic information may include information that has no particular correlation. For example, the subject characteristic information may be the subject's credit card number, personal information, demographic information, delinquency information, usage information, past reminder history, or a combination thereof. The subject characteristic information may include the history of past debt collection, credit-related information, or other information. In this embodiment, the case where the subject characteristic information is demographic information is taken as an example.

[0037] The effect is the success of the reminder. In other words, the effect is the achievement of the purpose for which the reminder is issued. For example, the effect is that the debt is collected by the reminder. The effect may be the amount collected by the reminder. The effect may vary depending on the scenario in which the estimation system 1 is used. The effect may be the ability to contact the subject rather than debt collection. For example, as a predetermined action, when contact (e.g., a telephone appointment) is made for a purpose other than a reminder, the effect is the ability to contact the subject. When an advertisement is presented as a predetermined action, the effect is that the subject views the advertisement, the subject selects the advertisement, or the subject purchases the product or service of the advertisement. When sales activities are carried out as a predetermined action, the effect is the success of the sales activities. For example, the effect is the ability to conclude a contract.

[0038] The probability estimation model M1 is a model for estimating the probability of obtaining an effect. For example, the probability estimation model M1 is a model for estimating the probability that a target person will make a payment when a reminder is made at a certain timing by a certain means. The probability estimation model M1 is a model created by a machine learning method. Various known machine learning methods can be used. The machine learning method may be any of supervised learning, semi-supervised learning, or unsupervised learning. For example, the probability estimation model M1 may be a neural network, a support vector machine, a large language model, or other models. There are various theories about the meaning of the term "machine learning", but the machine learning in this embodiment includes various known meanings. For example, the machine learning in this embodiment also includes AI (Artificial Intelligence) and deep learning in the inclusive sense. These meanings of machine learning are the same for other models (for example, the initial value estimation model M2 described later) other than the probability estimation model M1.

[0039] For example, the probability estimation model storage unit 101 stores a learned probability estimation model M1 in which the relationship between the target person characteristic information and the probability of obtaining an effect is learned. In this embodiment, the case where the server 10 performs the learning of the probability estimation model M1 is taken as an example, but other computers other than the server 10 may perform the learning of the probability estimation model M1. For example, the person-in-charge terminal 20 may perform the learning of the probability estimation model M1. The probability estimation model M1 includes a program that performs a series of processes such as calculation of embedded expressions and parameters referred to by the program. The parameters of the probability estimation model M1 are adjusted by learning. The program and parameters of the probability estimation model M1 may be the same as known programs and parameters. For example, the program of the probability estimation model M1 may include a plurality of layers such as an input layer that receives data input, an intermediate layer that performs calculations such as calculation of embedded expressions, and an output layer that outputs an estimation result. The parameters of the probability estimation model M1 may be weights and biases.

[0040] FIG. 4 is a diagram showing an example of the relationship between the input and output of the probability estimation model M1. The data format of each of the input and output of the probability estimation model M1 is basically the same during estimation and learning. The numerical value of the output in FIG. 4 is the probability of obtaining an effect. In the present embodiment, the case where the probability estimation model M1 is prepared for each communication means is taken as an example. For example, if three communication means such as person-to-person power supply, robot power supply, and SMS are prepared, as shown in FIG. 4, three probability estimation models M1, namely, the probability estimation model M1 for person-to-person power supply, the probability estimation model M1 for robot power supply, and the probability estimation model M1 for SMS, are prepared.

[0041] For example, the data storage unit 100 stores a training database in which training data necessary for learning the probability estimation model M1 is stored. The training data includes an input part that is input to the probability estimation model M1 during learning and an output part that becomes the correct answer during learning. The input part of the training data is basically in the same format as the input data input to the probability estimation model M1 during estimation. The output part of the training data is basically in the same format as the output data output from the probability estimation model M1 during estimation. Note that the input part of the training data may be in a format that is slightly different from the input data input to the probability estimation model M1 during estimation. Similarly, the output part of the training data may be in a format that is slightly different from the output data output from the probability estimation model M1 during estimation.

[0042] For example, the input part of the training data includes target person characteristic information for training. In the present embodiment, the input part of the training data also includes the timing at which a target person for training is urged. The output part of the training data indicates the probability (the probability that becomes the correct answer during learning) of obtaining an effect from the target person of the characteristics indicated by the target person characteristic information that is the input part of the training data. The training data may be created by the administrator of the estimation system 1 or may be created by a known tool. The server 10 performs learning of the probability estimation model M1 so that the output part of the training data is output when the input part of the training data is input. The server 10 performs learning of the probability estimation model M1 by adjusting the parameters of the probability estimation model M1 based on the training data.

[0043] Note that the algorithm for training the probability estimation model M1 may be a known algorithm used in the field of machine learning. For example, the server 10 may cause the probability estimation model M1 to learn training data based on an algorithm such as the gradient descent method or the error backpropagation method. The loss function used during learning may also be a known loss function. The server 10 calculates the loss, which is the error between the output part of the training data and the output from the probability estimation model M1 during learning, based on the loss function. The server 10 completes the learning when the loss becomes small to a certain extent. When the learning is completed, the server 10 records the learned probability estimation model M1 in the probability estimation model storage unit 101.

[0044] As shown in FIG. 4, when the probability estimation model M1 is prepared for each communication means, in the probability estimation model M1 of a certain communication means, the relationship between the target person characteristic information of the target person for training targeted at the communication means and the probability obtained from the target person for training is learned. For this reason, for each communication means, the training data that the probability estimation model M1 of the communication means learns is prepared. The data format of the training data that the probability estimation model M1 of each communication means learns may be the same as the data format of the training data that the probability estimation model M1 of other communication means learns. In the present embodiment, since three probability estimation models M1 such as the probability estimation model M1 for human power supply, the probability estimation model M1 for robot power supply, and the probability estimation model M1 for SMS are prepared, the server 10 causes the probability estimation model M1 of each of these three communication means to learn based on the training data of each communication means.

[0045] [Effect information acquisition unit] The effect information acquisition unit 102 acquires effect information regarding the effects obtained when a predetermined action is performed on each of a plurality of target persons by a predetermined means at a predetermined timing. In the present embodiment, since the predetermined means is a communication means for reminder and the predetermined action is a reminder, the effect information acquisition unit 102 acquires effect information regarding the effects obtained when a reminder is made to each of the plurality of target persons by a predetermined communication means at a predetermined timing.

[0046] The effect information is information that has some relation to the effects obtained from the target persons. The effect information may be a binary value indicating the presence or absence of an effect, but in the present embodiment, an example is given where the effect information is not a binary value but indicates the effects expected from the target persons by a numerical value of three or more levels. For example, the effect information acquisition unit 102 acquires the effect information of each of the plurality of target persons based on the probability of obtaining an effect when a reminder is made to each of the plurality of target persons by a predetermined communication means at a predetermined timing and the degree of the effect obtained from the target person. Note that the effect information may be expressed in other forms such as characters or symbols instead of numerical values.

[0047] The degree of the effect is the magnitude of the effect. In the present embodiment, when the collection of the claim from a target person with a large amount of arrears is successful, the effect of claim collection becomes large, so an example is given where the amount of arrears corresponds to the degree of the effect. The degree of the effect may vary depending on the scene where the estimation system 1 is used. For example, when communication (for example, telephone appointment) is made for a purpose other than reminder as a predetermined action, the effect is the sales or profit obtained from the target person by the communication. When an advertisement is presented as a predetermined action, the advertisement is the price of the product or service targeted by the advertisement. When business is conducted as a predetermined action, the effect is the sales or profit obtained by the business.

[0048] FIG. 5 is a diagram showing an example of the processing of each of the effect information acquisition unit 102, the cost information acquisition unit 103, the constraint information acquisition unit 104, and the estimation unit 105. In the present embodiment, the effect information acquisition unit 102 acquires the probability that an effect can be obtained from each of a plurality of target persons based on the effect information of each of the plurality of target persons and the probability estimation model M1, and acquires effect information based on the probability of each of the plurality of target persons. For example, the effect information acquisition unit 102 acquires the target person characteristic information of each of the plurality of target persons. In the present embodiment, since the case where demographic information is used as the target person characteristic information is taken as an example, the effect information acquisition unit 102 acquires the demographic information of each of the plurality of target persons from the target person database DB as the target person characteristic information of the target person. The effect information acquisition unit 102 may acquire the target person characteristic information of each of the plurality of target persons from another database other than the target person database DB, another computer other than the server 10, or an external information storage medium.

[0049] In the present embodiment, it is assumed that a planned period in which debt collection is planned is predetermined. The effect information acquisition unit 102 identifies a plurality of timings within the planned period. In the example of FIG. 5, it is assumed that the planned period is from February 5, 2024 to February 9, 2024. Therefore, the effect information acquisition unit 102 identifies the five days of the planned period as a plurality of timings. The planned period may be of any length. For example, the planned period may be several days, one week, two weeks, or one month. It is assumed that the data indicating the planned period is stored in the data storage unit 100. The planned period may be specified by a person in charge or an administrator.

[0050] For example, the effect information acquisition unit 102 inputs, for each target person, input data including the target person's characteristic information and the timing within the planned period into the probability estimation model M1. In the example of FIG. 5, since there are five timings, there are five input data for each target person. In the present embodiment, since the probability estimation model M1 is prepared for each communication means, the effect information acquisition unit 102 inputs the input data to each of the plurality of probability estimation models M1. In the example of FIG. 5, the effect information acquisition unit 102 inputs the input data of the target person to each of the probability estimation model M1 for human power supply, the probability estimation model M1 for robot power supply, and the probability estimation model M1 for SMS for each target person.

[0051] For example, when the input data is input, the probability estimation model M1 calculates the embedded representation of the input data based on the parameters adjusted by learning. The embedded representation is information indicating the characteristics of the input data. The embedded representation may also be called a feature quantity. The embedded representation may be in any form, for example, vector form, array form, matrix form, a plurality of numerical values, a single numerical value, or other forms. If the probability estimation model M1 is a large language model, the probability estimation model M1 divides the input data into tokens and then calculates the embedded representation of each individual token. The probability estimation model M1 outputs a probability corresponding to the embedded representation calculated from the input data.

[0052] In the present embodiment, since the probability estimation model M1 is prepared for each communication means, the effect information acquisition unit 102 acquires the probability for each communication means based on the probability estimation model M1 of the communication means. The processes executed by the individual probability estimation models M1 may be the same. Since the parameters of the individual probability estimation models M1 become different values by learning, even if the input data is the same, different output data are acquired. In the example of FIG. 5, the effect information acquisition unit 102 inputs the input data of the target person to each of the probability estimation model M1 for human power supply, the probability estimation model M1 for robot power supply, and the probability estimation model M1 for SMS for each target person, and acquires the probability indicated by the output data output from each of these three probability estimation models M1.

[0053] For example, the effect information acquisition unit 102 may calculate, for each target person, the probability that an effect can be obtained when a reminder is made to the target person by each of a plurality of communication means, and calculate an expected recovery amount which is a value obtained by multiplying the overdue amount indicating the degree of the effect by the probability. The expected recovery amount is an example of effect information. The expected recovery amount can also be said to be the degree to which an effect is expected. In the example of FIG. 5, the overdue amount of target person A is 100,000 yen. The effect information acquisition unit 102 calculates the expected recovery amount by multiplying the overdue amount of 100,000 yen of target person A by the probability that an effect can be obtained when a reminder is made by each of the three communication means. The overdue amount of target person B is 200,000 yen. The effect information acquisition unit 102 calculates the expected recovery amount by multiplying the overdue amount of 200,000 yen of target person B by the probability that an effect can be obtained when a reminder is made by each of the three communication means.

[0054] Note that the effect information acquired by the effect information acquisition unit 102 is not limited to the examples of the present embodiment. For example, the effect information acquisition unit 102 may calculate the expected recovery amount by multiplying the probability estimated by the probability estimation model M1 by a weight coefficient and then multiplying by the overdue amount, and acquire it as effect information. The effect information acquisition unit 102 may calculate the expected recovery amount by multiplying the probability estimated by the probability estimation model M1 by the overdue amount multiplied by the weight coefficient, and acquire it as effect information. The effect information acquisition unit 102 may acquire the probability rather than the expected recovery amount as effect information. The effect information acquisition unit 102 may acquire the overdue amount rather than the expected recovery amount as effect information. In this case, the effect information acquisition unit 102 may not calculate the probability. The server 10 may not include the probability estimation model storage unit 101.

[0055] [Cost Information Acquisition Unit] The cost information acquisition unit 103 acquires cost information regarding the cost required when a reminder is made to each of a plurality of target persons by a predetermined communication means at a predetermined timing. The cost information is information indicating the degree of cost. The cost information may indicate a monetary cost, a time cost, a hardware or software resource cost, or other costs. For example, the cost information indicates a numerical value of the cost. The cost information may be expressed by characters or symbols in addition to numerical values. In this embodiment, a case where the cost information is stored in the data storage unit 100 is taken as an example. Therefore, the cost information acquisition unit 103 acquires the cost information from the data storage unit 100. The cost information may be stored in another computer other than the server 10 or an external information storage medium. In this case, the cost information acquisition unit 103 may acquire the cost information from another computer or an external information storage medium.

[0056] In this embodiment, it is assumed that the cost varies depending on the communication means. Further, it is assumed that the cost of each individual communication means is a fixed value. That is, a case is taken as an example where the cost is not changed by other elements other than the communication means and the cost of the communication means is predetermined. The mode in which the cost changes due to other elements will be described in a modified example described later. In this embodiment, the data storage unit 100 stores the cost information of each of a plurality of communication means. The cost information acquisition unit 103 acquires the cost information of each of a plurality of communication means from the data storage unit 100. When there is only one communication means, the data storage unit 100 stores the cost information of the one communication means. The cost information acquisition unit 103 may acquire the cost information of the one communication means. Even when there are a plurality of communication means, the cost information may be common among the plurality of communication means.

[0057] In the example of FIG. 5, since three communication means, namely human operator call, robot call, and SMS, are provided, the cost information acquisition unit 103 acquires the cost information of human operator call, the cost information of robot call, and the cost information of SMS. The cost information of human operator call indicates the cost when a human operator call is made. For example, the cost information of human operator call indicates the total of labor costs and call charges for a human operator call. The cost information of robot call indicates the cost when a robot call is made. For example, the cost information of robot call indicates the call charges for a robot call. The cost information of SMS indicates the cost when an SMS is sent. For example, the cost information of SMS indicates the usage fee for SMS. The cost information may indicate the cost per communication using these communication means, or may indicate the total cost of multiple communications.

[0058] [Restriction Information Acquisition Unit] The restriction information acquisition unit 104 acquires restriction information regarding the cost restrictions allowed for reminders. The cost restriction is the upper limit value of the cost. The cost restriction can also be referred to as the allowable range of the cost. The cost restriction may be a restriction at each individual timing, or may be a restriction for the entire planned period during which debt collection is planned. In this embodiment, an example is given where the restriction information is stored in the data storage unit 100. Therefore, the restriction information acquisition unit 104 acquires the restriction information from the data storage unit 100. The restriction information may be stored in another computer other than the server 10 or an external information storage medium. In this case, the restriction information acquisition unit 104 may acquire the restriction information from another computer or an external information storage medium.

[0059] In this embodiment, it is assumed that the restrictions are different depending on the communication means. Furthermore, it is assumed that the restrictions for each individual communication means are fixed values. That is, taking the case where the restriction is one-to-one with the communication means without the restriction changing due to other elements than the communication means as an example. The mode in which the restriction changes due to other elements will be described in a modification example described later. In this embodiment, the data storage unit 100 stores the restriction information for each of the plurality of communication means. The restriction information acquisition unit 104 acquires the restriction information in which the restriction corresponding to the communication means is defined. That is, the restriction information acquisition unit 104 acquires the restriction information for each of the plurality of communication means from the data storage unit 100. When there is only one communication means, the data storage unit 100 stores the restriction information for the one communication means. The restriction information acquisition unit 104 may acquire the restriction information for the one communication means. Even when there are a plurality of communication means, the restriction information may be common among the plurality of communication means.

[0060] In the example of FIG. 5, since three communication means of human-powered call, robot-powered call, and SMS are prepared, the restriction information acquisition unit 104 acquires the restriction information for human-powered call, the restriction information for robot-powered call, and the restriction information for SMS. The restriction information for human-powered call indicates the restriction when a human-powered call is made. For example, the restriction information for human-powered call indicates the amount of money allowed for labor costs and call charges due to human-powered call. The restriction information for human-powered call may indicate the upper limit of the number of calls allowed for human-powered call (for example, the upper limit of resources of hardware or software used for human-powered call). The restriction information for robot-powered call indicates the restriction when a robot-powered call is made. For example, the restriction information for robot-powered call indicates the amount of money allowed for call charges due to robot-powered call. The restriction information for robot-powered call may indicate the upper limit of the number of calls allowed for robot-powered call (for example, the upper limit of resources of hardware or software used for robot-powered call). The restriction information for SMS indicates the restriction when SMS is sent. For example, the restriction information for SMS indicates the amount of money allowed for the usage fee of SMS. The restriction information for SMS may indicate the upper limit of the number of messages allowed for SMS.

[0061] [Estimation Unit] Based on the effect information and cost information of each of the plurality of target persons and the constraint information, the estimation unit 105 estimates a combination of a target person to be urged, at least one of a predetermined timing and a predetermined communication means, so that the effect can be obtained efficiently as a whole within the range of the constraint. In the present embodiment, a case where the estimation unit 105 estimates a combination of a target person, a predetermined timing, and a predetermined communication means will be described. However, the estimation unit 105 may estimate a combination of a target person and a predetermined timing without estimating the predetermined communication means. The estimation unit 105 may estimate a combination of a target person and a predetermined communication means without estimating the predetermined timing. As described above, in the present embodiment, the combination is a combination of reminders.

[0062] The range of the constraint means that the cost is equal to or less than the cost indicated by the constraint. That is, the range of the constraint means that the cost does not exceed the cost indicated by the constraint. The fact that the total cost corresponding to the reminder combination estimated by the estimation unit 105 is equal to or less than the cost indicated by the constraint corresponds to being within the range of the constraint. The whole means the whole of the plurality of target persons. That is, the whole means the whole of the credit card service. Obtaining the effect efficiently means that the overall effect is relatively high. In other words, the fact that the effect on the cost required for the reminder is high corresponds to obtaining the effect efficiently. That is, obtaining the effect efficiently means that the cost performance in the reminder is good. Obtaining the effect efficiently means that the cost-effectiveness is high.

[0063] In the example of FIG. 5, the variable being 1 corresponds to a reminder being issued. The variable being 0 corresponds to no reminder being issued. For example, there is a variable for each combination of the person to be reminded, the timing, and the means of communication. Let the number of persons to be reminded be x (x is a natural number), the number of timings be y (y is a natural number), and the number of means of communication be z (z is a natural number). Then, there are x × y × z number of variables. In the example of FIG. 5, assume that the timing is for 5 days from February 5, 2024 to February 9, 2024. The means of communication are three: personal call, robot call, and SMS. Therefore, there are 5 × 3 = 15 variables per person to be reminded. If there are 1000 persons to be reminded, there are 1000 × 15 = 15000 variables. In the example of FIG. 5, two variables are 1. For example, the first variable that is 1 indicates that a robot call is made to person A on February 5, 2024. The second variable that is 1 indicates that an SMS is sent to person B on February 5, 2024.

[0064] For example, the estimation unit 105 determines at least one combination among the candidates for the combination where a reminder is issued as the combination where a reminder is actually issued. In the example of FIG. 5, the estimation unit 105 determines the combination of 0 and 1 of the variables. In the above-mentioned example, the estimation unit 105 determines the combination of 0 and 1 of the 15000 variables. Since the combination of reminders where the variable is 0 is not adopted, no cost is incurred, but since the combination of reminders where the variable is 1 is adopted, cost is incurred. The estimation unit 105 estimates the combination of reminders such that the total value of the costs of the combination of reminders where the variable is 1 is within the range of the constraint and the total value of the expected recovery efficiency of the combination of reminders where the variable is 1 is relatively higher than that of other combinations.

[0065] For example, the estimation unit 105 estimates the combination for which a reminder is to be issued based on a metaheuristic method. The metaheuristic method can also be said to be a method for solving a combinatorial optimization problem. Combinatorial optimization does not necessarily require a precise solution to be identified; it suffices to identify an appropriate solution to a certain extent. For example, the metaheuristic method may be a genetic algorithm, particle swarm optimization, ant colony algorithm, or simulated annealing. In this embodiment, an example is given where the estimation unit 105 determines the combination for which a reminder is to be issued based on a genetic algorithm, but the estimation unit 105 may estimate the combination for which a reminder is to be issued based on other metaheuristic methods.

[0066] In this embodiment, the estimation unit 105 calculates an index related to the efficiency of the effects obtained from each of a plurality of subjects based on the effect information and cost information of each of the plurality of subjects, and estimates the combination of reminders based on the index of each of the plurality of subjects. In the example of FIG. 5, the estimation unit 105 calculates the expected recovery efficiency of each subject based on, for each subject, the expected recovery amount, which is an example of the effect information of the subject, and the expected cost, which is an example of the cost information of the subject. For example, the estimation unit 105 calculates the expected recovery efficiency of each subject by dividing the expected recovery amount of the subject by the expected cost of the subject. The estimation unit 105 calculates the expected recovery efficiency for each combination of reminders.

[0067] The expected recovery efficiency is an example of an index related to efficiency. The index related to efficiency can be said to be the degree of effect per unit cost. The index related to efficiency may be any index calculated by substituting the effect information and cost information into a calculation formula. The index related to efficiency is not limited to the expected recovery efficiency. For example, the estimation unit 105 may not simply divide the expected recovery amount of each subject by the expected cost of the subject, but may divide after multiplying the expected recovery amount or expected cost by a coefficient corresponding to the subject.

[0068] In the present embodiment, since restrictions corresponding to the communication means are defined, the estimation unit 105 estimates a combination of reminders within the range of the restrictions corresponding to the communication means. For example, for each communication means, the estimation unit 105 may estimate a combination of the person on whom an action is to be performed, at least one of the timing and the communication means, within the range of the restrictions corresponding to the communication means so that the effect can be obtained efficiently as a whole. The estimation unit 105 estimates a combination of reminders such that, for each communication means, the total value of the costs of the combination of reminders with the variable being 1 is within the range of the restrictions of the communication means, and the total value of the expected recovery efficiency of the combination of reminders with the variable being 1 is relatively high. Also in this case, the estimation unit 105 may estimate a combination of reminders based on the meta-heuristic method described above.

[0069] Note that the estimation unit 105 may estimate a combination of reminders such that, for a certain person, there is only one combination of reminders with the variable being 1. When the total cost of the combination of reminders with the variable being 1 for all persons or a certain number of persons is within the range of the restrictions and there is a margin up to the upper limit of the restrictions, the estimation unit 105 may estimate a combination of reminders such that, for a certain person, there are two or more combinations of reminders with the variable being 1. The estimation unit 105 may estimate a combination of reminders that can obtain an effect efficiently as a whole based on a method other than the meta-heuristic method (for example, a machine learning method).

[0070] [4. Processing Executed by the Estimation System] FIG. 6 is a diagram showing an example of the processing executed by the estimation system 1. In the present embodiment, the case where the main processing of the estimation system 1 is executed by the server 10 is taken as an example. The processing of FIG. 6 is executed by the control unit 11 executing the program stored in the storage unit 12. It is assumed that the learning of the probability estimation model M1 has been completed when the processing of FIG. 6 is executed.

[0071] As shown in FIG. 6, the server 10 acquires the subject characteristic information of each of a plurality of subjects based on the subject database DB (S1). For each subject, the server 10 calculates, based on the subject characteristic information of the subject, each of a plurality of timings, and the probability estimation model M1 of each of a plurality of communication means, the probability that an effect can be obtained from the subject when a reminder is made by the communication means at the timing for the subject (S2).

[0072] For each subject, the server 10 calculates the expected recovery amount of the subject based on the overdue amount indicated by the overdue amount of the subject stored in the subject database DB and the probability of the subject calculated in S2, and acquires it as the effect information of the subject (S3). For each communication means, the server 10 acquires cost information indicating the expected cost of the communication means stored in the storage unit 12 (S4). For each subject, the server 10 calculates the expected recovery efficiency based on the effect information of the subject acquired in S3 and the cost information acquired in S4, and acquires it as an index related to efficiency (S5). By the processes of S1 to S5, the calculation of information other than the variables shown in FIG. 5 is completed.

[0073] For each communication means, the server 10 acquires the constraint condition information of the communication means stored in the storage unit 12 (S6). Based on the index acquired in S5 and the constraint condition information acquired in S6, the server 10 estimates the combination of reminders so that an effect can be obtained efficiently as a whole within the range of constraints of each of the plurality of communication means (S7). In S7, based on the above-described metaheuristic method, the server 10 estimates the combination of reminders such that the total value of the costs of the combination of reminders with the variable being 1 is within the range of the constraints of the communication means, and the total value of the expected recovery efficiency of the combination of reminders with the variable being 1 is relatively higher than that of other combinations. The server 10 transmits reminder combination data indicating the combination of reminders estimated in S8 to the operator terminal 20 (S8), and this process ends.

[0074] FIG. 7 is a diagram showing an example of a screen displayed on the person-in-charge terminal 20. For example, when the person-in-charge terminal 20 receives the reminder combination data transmitted by the server 10 at S8, the person-in-charge terminal 20 causes the display unit 25 to display the screen SC of FIG. 7 based on the reminder combination data. The combination data may show not only the combination of reminders for which the variable has become 1 but also the total value of the expected recovery efficiency obtained by the reminders. The total value is calculated by the server 10. The person in charge grasps the plan for collecting his or her own claims by checking the screen SC. The person in charge may modify the plan displayed on the screen SC by an operation from the operation unit 24.

[0075] [Summary of Embodiment] The estimation system 1 of the present embodiment acquires the effect information of each of a plurality of target persons. The estimation system 1 acquires the cost information of a plurality of target persons. The estimation system 1 acquires constraint information. Based on the effect information and cost information of each of the plurality of target persons and the constraint information, the estimation system 1 estimates a combination of a target person to be reminded, at least one of timing and communication means, so that an effect can be efficiently obtained as a whole within the range of the constraint. Thereby, the estimation system 1 can estimate the combination of reminders so that the effect can be efficiently obtained as a whole while satisfying the cost constraint, and thus the reminder to the target person can be made more efficient. For example, the estimation system 1 can realize reminders with high cost performance (cost vs. effect) as a whole, rather than only succeeding in collecting from a certain specific target person. The estimation system can save the trouble of the person in charge and the administrator making a reminder plan, and thus can improve the convenience of the person in charge and the administrator.

[0076] In addition, for each of a plurality of target persons, the estimation system 1 acquires the effect information of the target person based on the probability of obtaining an effect when a reminder is given by a predetermined communication means at a predetermined timing and the degree of the effect obtained from the target person. As a result, the estimation system 1 can estimate the combination of reminders based on the effect information in which the probability of obtaining an effect and the degree of the effect are comprehensively considered, so that the estimation accuracy of the combination of reminders can be improved. Consequently, the estimation system 1 can make the reminder to the target person more efficient.

[0077] Furthermore, the estimation system 1 calculates an index related to the efficiency of the effect obtained from each of the plurality of target persons based on the effect information and cost information of each of the plurality of target persons, and estimates the combination of reminders based on the index of each of the plurality of target persons. As a result, the estimation system 1 can estimate the combination of reminders based on the index related to the efficiency of the effect obtained from each of the plurality of target persons, so that the estimation accuracy of the combination of reminders can be improved. For example, the estimation system 1 can make it easier to apply a meta-heuristic method by means of the index. Consequently, the estimation system 1 can make the reminder to the target person more efficient.

[0078] In addition, the estimation system 1 acquires constraint information in which constraints corresponding to the communication means are defined. The estimation system 1 estimates the combination of reminders within the range of the constraints corresponding to the communication means. As a result, even if the constraints are different for each communication means, the estimation system 1 can estimate the combination of reminders within the achievable range by estimating a combination that can efficiently obtain an effect as a whole within the range of the constraints of each of the plurality of communication means. Since the person in charge can give a reminder by efficiently combining a plurality of communication means, the estimation system 1 can make the reminder to the target person more efficient.

[0079] Further, the estimation system 1 obtains the probability of obtaining an effect from each of a plurality of target persons based on the effect information of each of the plurality of target persons and the probability estimation model M1, and obtains the effect information based on the probability of each of the plurality of target persons. Thereby, the estimation system 1 can improve the estimation accuracy of the probability of obtaining an effect from each of the plurality of target persons by using the probability estimation model M1. As a result, since the estimation accuracy of the combination of reminders is improved, the estimation system 1 can make the reminder to the target person more efficient.

[0080] Also, each of the plurality of target persons is a delinquent who has defaulted on the payment of a credit card. The predetermined means is a communication means for reminding the delinquent. The predetermined action is a reminder. The cost is the cost for the reminder. The estimation system 1 can make the reminder to the delinquent who has defaulted on the payment of a credit card more efficient. The estimation system 1 can achieve efficient debt collection as a whole for the credit card service.

[0081] [6. Modification Example] The present disclosure is not limited to the embodiments described above. The present disclosure can be appropriately changed without departing from the spirit of the present disclosure.

[0082] FIG. 8 is a diagram showing an example of functions realized in a modification example. For example, the server 10 includes a future risk information acquisition unit 106 and an initial value estimation model storage unit 107. The future risk information acquisition unit 106 is realized by the control unit 11. The initial value estimation model storage unit 107 is realized by the storage unit 12.

[0083] [6-1. Modification Example 1] For example, depending on the means of communication, the cost may vary depending on whether or not the target person responds. When the means of communication is a person-to-person call, if the target person responds to the call from the person in charge, it takes time for the person in charge to talk to the target person, so the cost becomes high. If the target person does not respond to the call from the person in charge, it does not take time for the person in charge to talk to the target person, so the cost becomes low. Therefore, since the cost is not a fixed value as in the embodiment, the cost may be calculated based on the probability of response by the target person.

[0084] The cost information acquisition unit 103 of Modification 1 calculates the probability of response when a reminder is made to each of a plurality of target persons by a predetermined means of communication at a predetermined timing, and acquires cost information based on the probability of each of the plurality of target persons. In Modification 1, the case where the probability of response is calculated by the probability estimation model M1 is taken as an example, but the probability of response may be calculated by a model different from the probability estimation model M1. The output data from the probability estimation model M1 of Modification 1 indicates the probability of response, not the probability of payment. Therefore, the output part of the training data also indicates the probability of response, not the probability of payment by the training target person.

[0085] For example, the cost information acquisition unit 103 inputs input data indicating a combination of the target person's characteristic information of the target person and a predetermined timing to the probability estimation model M1 for each target person. The input data may be the same as in the embodiment. The probability estimation model M1 calculates the embedded representation of the input data and outputs output data indicating the probability of response according to the embedded representation. The cost information acquisition unit 103 acquires the probability of response of the target person by acquiring the output data. Although it is different from the embodiment in that the output data indicates the probability of response, the processing executed by the probability estimation model M1 may be the same as in the embodiment.

[0086] For example, the cost information acquisition unit 103 calculates the cost of each of a plurality of target persons such that the higher the probability of the response of each target person, the higher the final cost required for the target person, and acquires cost information. The cost information acquisition unit 103 acquires cost information indicating the final cost required for each target person by substituting the probability of the target person and the cost of the communication means into a predetermined calculation formula for each target person. The relationship between probability and cost is defined in the calculation formula. The cost information acquisition unit 103 may acquire cost information indicating the final cost required for each target person by multiplying the probability of the target person and the cost of the communication means for each target person. The cost information acquisition unit 103 may perform such multiplication based on a predetermined weight coefficient. Although the method of acquiring cost information is different from that of the embodiment, other processes may be the same as those of the embodiment.

[0087] The estimation system 1 of Modification 1 calculates the probability of response when a reminder is given to each of a plurality of target persons by a predetermined communication means at a predetermined timing, and acquires cost information based on the probability of each of the plurality of target persons. Thereby, since the estimation system 1 can estimate the combination of reminders based on the cost information corresponding to the probability of the response of the target person, the estimation accuracy of the combination of reminders can be improved. As a result, the estimation system 1 can make the reminder to the target person more efficient.

[0088] [6-2. Modification 2] For example, even if the person in charge contacts the target person and the target person does not make a payment, the person in charge may give a reminder to the target person again. In this case, since the person in charge will give a reminder a plurality of times, the total cost will increase when future costs are also considered. For this reason, the cost may be calculated based on the probability of obtaining an effect from the target person.

[0089] The cost information acquisition unit 103 of Modification 2 calculates the probability that an effect can be obtained when a reminder is given to each of a plurality of target persons by a predetermined communication means at a predetermined timing, and acquires cost information based on the probability of each of the plurality of target persons. For example, the cost information acquisition unit 103 calculates the probability of obtaining an effect from the target person based on the probability estimation model M1 described in the embodiment. The method of calculating the probability may be the same as that in the embodiment.

[0090] For example, the lower the probability of obtaining an effect from each of the plurality of target persons, the higher the final cost required for the target person. The cost information acquisition unit 103 calculates the cost of the target person and acquires cost information. The cost information acquisition unit 103 substitutes the probability of obtaining an effect from the target person and the cost of the communication means into a predetermined calculation formula for each target person, thereby obtaining cost information indicating the final cost required for the target person. The relationship between probability and cost is defined in the calculation formula. The cost information acquisition unit 103 may obtain cost information indicating the final cost required for the target person by multiplying the value obtained by subtracting the probability of obtaining an effect from the target person from a predetermined value (for example, 1) and the cost of the communication means for each target person. The cost information acquisition unit 103 may perform such multiplication based on a predetermined weighting factor. Although the method of acquiring cost information is different from that in the embodiment, other processes may be the same as those in the embodiment.

[0091] The estimation system 1 of Modification 2 calculates the probability that an effect can be obtained when a reminder is given to each of a plurality of target persons by a predetermined communication means at a predetermined timing, and acquires cost information based on the probability of each of the plurality of target persons. Thereby, the estimation system 1 can estimate the combination of reminders based on the cost information corresponding to the probability of obtaining an effect from the target person, so that the estimation accuracy of the combination of reminders can be improved. As a result, the estimation system 1 can make the reminder to the target person more efficient.

[0092] [6-3. Modification 3] For example, some of the subjects may not only have past overdue payments but also have applied for a large number of installment payments or bonus payments. Although no overdue payments have occurred at present for future payments in installment payments or bonus payments, there is a possibility that the subject may delay future payments in installment payments or bonus payments. Therefore, future payments in installment payments or bonus payments can be regarded as future risks. In Modification 3, an example is given where such future risks are considered and the combination of reminders is determined.

[0093] The estimation system 1 of Modification 3 includes a future risk information acquisition unit 106. The future risk information acquisition unit 106 acquires future risk information regarding the risk that reminders will be necessary in the future for each of a plurality of subjects. The future risk information is information related to the necessity of reminders. The future risk information may be information directly indicating the risk or information having a correlation with the risk. In Modification 3, an example is given where the future risk information is represented by a numerical value, but the future risk information may also be represented by characters or symbols other than numerical values.

[0094] For example, the future risk information indicates the amount of the subject's installment payment or bonus payment. Assume that the amount of the subject's installment payment or bonus payment is shown in the usage information stored in the subject database DB. Therefore, the future risk information acquisition unit 106 acquires the future risk information by acquiring the usage information of the subject from the subject database DB. The future risk information acquisition unit 106 may also acquire the future risk information from another database other than the subject database DB, another computer other than the server 10, or an external information storage medium.

[0095] Note that the future risk information may indicate, in addition to the amount of installment payments or bonus payments for the target person, other amounts for which delinquency by the target person may occur in the future. For example, the future risk information may indicate the cashing amount of the target person. When a model for estimating future risks is prepared, the future risk information may be estimated based on the target person's target person characteristic information and the model. The future risk information acquisition unit 106 may acquire future risk information using a known model such as a credit model. The future risk information acquisition unit 106 may use the credit information stored by an external credit institution as future risk information.

[0096] The estimation unit 105 of Modification Example 3 estimates the combination of reminders based on the future risk information as well. For example, the estimation unit 105 multiplies the expected recovery efficiency based on a weight coefficient corresponding to the future risk information. For example, the higher the future risk indicated by the future risk information, the larger the weight coefficient. The estimation unit 105 calculates the expected recovery efficiency of a target person so that the higher the future risk of the target person, the larger the expected recovery efficiency of the target person, making it easier for the target person to be selected as a reminder target.

[0097] The estimation system 1 of Modification Example 3 acquires the future risk information of each of a plurality of target persons. The estimation system 1 estimates the combination of reminders based on the future risk information as well. Thereby, since the estimation system 1 can estimate the combination of reminders in consideration of the future risks of the target persons, the estimation accuracy of the combination of reminders can be improved. As a result, the estimation system 1 can make the reminders to the target persons more efficient.

[0098] [6-4. Modification Example 4] For example, in Modification 3, the future risk information acquisition unit 106 may acquire future risk information based on the probability of a risk occurring for each of a plurality of target persons and the degree of the risk. In Modification 4, the future risk information acquisition unit 106 calculates the probability of a risk occurring based on the demographic information of each of the plurality of target persons. For example, the future risk information acquisition unit 106 inputs the demographic information of a target person to a credit model that calculates the probability of a risk occurring, and acquires the probability output by the credit model. The probability may be calculated based on a model other than the credit model. The probability may be calculated by a program for calculating the probability, especially not a machine learning model.

[0099] The degree of the risk is the magnitude of the risk. For example, the amount of installment payment or bonus payment of a target person corresponds to the degree of the risk. The degree of the risk may be the cashing amount of the target person. Assume that the degree of the risk is stored in the target person database DB. The future risk information acquisition unit 106 acquires the degree of the risk of the target person from the target person database DB. The future risk information acquisition unit 106 may acquire the degree of the risk of the target person from another database other than the target person database DB, another computer other than the server 10, or an external information storage medium.

[0100] For example, the future risk information acquisition unit 106 may acquire the future risk information of a target person as a numerical value obtained by multiplying the probability of a risk occurring for each of a plurality of target persons by the degree of the risk of the target person. The future risk information acquisition unit 106 may acquire a numerical value obtained by multiplying these based on a weight coefficient associated with at least one of the probability of a risk occurring and the degree of the risk as the future risk information. The future risk information acquisition unit 106 may input the probability of a risk occurring for each of the plurality of target persons and the degree of the risk of the target person to a model for estimating future risk information, and acquire the future risk information of the target person output by the model.

[0101] The estimation system 1 of Modification Example 4 acquires future risk information based on the probability that a risk occurs for each of a plurality of target persons and the degree of the risk. Thereby, the estimation system 1 can improve the estimation accuracy of the future risk information. As a result, since the estimation system 1 can improve the estimation accuracy of the combination of reminders, the reminders for the target persons can be made more efficient.

[0102] [6-5. Modification Example 5] For example, the method for acquiring the future risk information is not limited to the examples of Modification Examples 3 and 4. For example, when each of a plurality of target persons uses a predetermined service, the future risk information acquisition unit 106 may acquire the future risk information based on the past usage status of the service by each of the plurality of target persons. The usage status is shown in the usage status information stored in the target person database DB. For example, the future risk information acquisition unit 106 determines whether or not the target person has made a past payment delinquency based on the past usage status of each of the plurality of target persons. The future risk information acquisition unit 106 may specify the past delinquency amount of the target person based on the past usage status of each of the plurality of target persons.

[0103] For example, the future risk information acquisition unit 106 acquires the future risk information so that the risk of a target person who has made a past payment delinquency is higher than the risk of a target person who has not made a past payment delinquency. Assume that the calculation formula required for acquiring the future risk information is stored in the data storage unit 100. The relationship between the presence or absence of delinquency and the risk is shown in the calculation formula. The future risk information acquisition unit 106 acquires the future risk information indicating the risk corresponding to the presence or absence of delinquency based on the calculation formula.

[0104] For example, the future risk information acquisition unit 106 may acquire future risk information such that the higher the amount of arrears that a certain subject has defaulted on in the past, the higher the risk of the subject. In this case, the calculation formula shows the relationship between the amount of arrears and the risk. The future risk information acquisition unit 106 acquires future risk information indicating the risk corresponding to the amount of arrears based on the calculation formula. The future risk information acquisition unit 106 may acquire future risk information based on a model that estimates future risk information based on the usage status by the subject.

[0105] The estimation system 1 of Modification 5 acquires future risk information based on the past usage status of the service by each of a plurality of subjects. Thereby, the estimation system 1 can improve the estimation accuracy of the future risk information. As a result, since the estimation system 1 can improve the estimation accuracy of the combination of reminders, the reminders for the subject can be made more efficient.

[0106] [6-6. Modification 6] For example, the estimation unit 105 may estimate the combination of reminders so that reminders are preferentially made for the subject with a relatively large effect. In the example of FIG. 5, the estimation unit 105 increases the weight coefficient used in the calculation of the expected recovery efficiency as the expected recovery amount increases. It is assumed that the relationship between the expected recovery amount and the weight coefficient is predetermined in the data storage unit 100. The estimation unit 105 may acquire the weight coefficient corresponding to the expected recovery amount based on the relationship. The estimation unit 105 calculates the expected recovery efficiency by dividing the value obtained by multiplying the expected recovery amount by the weight coefficient by the expected cost. After calculating the expected recovery efficiency in the same manner as in the embodiment, the estimation unit 105 may multiply by a weight coefficient corresponding to the amount of the expected recovery amount.

[0107] For example, the estimation unit 105 may increase the weight coefficient used in the calculation of the expected recovery efficiency as the amount of arrears of the target person increases. Assume that the relationship between the amount of arrears and the weight coefficient is predefined in the data storage unit 100. The estimation unit 105 may acquire a weight coefficient corresponding to the amount of arrears based on the relationship. The estimation unit 105 may calculate the expected recovery amount by multiplying the amount of arrears by the weight coefficient and then multiplying by the probability. Even if the estimation unit 105 performs such a calculation, it can estimate the combination of reminders so that reminders are preferentially made for target persons with relatively large effects.

[0108] The estimation system 1 of Modification 6 estimates the combination of reminders so that actions are preferentially taken for target persons with relatively large effects. Thereby, in the estimation system 1, it becomes easier to select target persons with relatively large effects, so that the effect obtained when the reminder for the target person is successful can be increased.

[0109] [6-7. Modification 7] For example, the constraint information acquisition unit 104 may acquire constraint information in which constraints corresponding to a predetermined timing are defined. Assume that the relationship between the timing and the constraints is stored in the data storage unit 100. In Modification 7, the data storage unit 100 stores the constraint information for each of a plurality of timings. The constraint information acquisition unit 104 acquires the constraint information for each of the plurality of timings from the data storage unit 100. When there is only one timing, the data storage unit 100 stores the constraint information for the one timing. The constraint information acquisition unit 104 may acquire the cost information for the one timing. Even when there are a plurality of timings, the cost information may be common for the plurality of timings.

[0110] The estimation unit 105 of Modification Example 7 estimates a combination of reminders within the range of constraints according to the timing. For example, the estimation unit 105 may estimate, for each timing, a combination of the person on whom an action is to be performed, and at least one of the timing and the communication means, so that the effect can be efficiently obtained as a whole within the range of constraints according to the timing. The constraints may include conditions other than cost. For example, the constraints may include the condition that the total value of the expected recovery efficiency is equal to or greater than a predetermined value. In this case, the satisfaction of the condition corresponds to being within the range of the constraints. The predetermined value may vary for each timing. For example, the predetermined value may be a value according to the recovery amount in the same period in the past. Although the method for obtaining the constraints used at the time of estimating the combination of reminders is different from that of the embodiment, the estimation of the combination of reminders itself may be the same as that of the embodiment.

[0111] The estimation system 1 of Modification Example 7 acquires constraint information in which constraints according to a predetermined timing are defined. The estimation system 1 estimates a combination within the range of constraints according to a predetermined timing. Thereby, even if the constraints are different for each timing, the estimation system 1 can estimate a combination that can efficiently obtain the effect as a whole within the range of constraints for each of a plurality of timings.

[0112] [6-8. Modification Example 8] For example, the estimation by the estimation unit 105 may end quickly or may take a long time depending on the combination used as the initial value. The initial value is the combination at the time of starting the estimation of the combination. In the example of FIG. 5, the initial value of the combination of 0 and 1 of the variables corresponds to the initial value of Modification Example 8. Therefore, if the estimation system 1 can improve the accuracy of the initial value used in the estimation by the estimation unit 105, it will be possible to quickly estimate an efficient combination of reminders as a whole. Therefore, in Modification Example 8, a case where a model for improving the accuracy of the initial value is prepared is taken as an example.

[0113] The estimation system 1 of Modification 8 includes an initial value estimation model storage unit 107. The initial value estimation model storage unit 107 stores an initial value estimation model M2 in which the relationship between the training effect information and cost information, and the combination of the training incentives has been learned. In Modification 8, similar to the embodiment, the case where the server 10 performs the learning of the initial value estimation model M2 is taken as an example, but other computers other than the server 10 may perform the learning of the initial value estimation model M2. For example, the person-in-charge terminal 20 may perform the learning of the initial value estimation model M2. The initial value estimation model M2 includes a program that performs calculations such as embedded expressions, and parameters referred to by the program. Among the initial value estimation models M2, the parameters are adjusted by learning. The program and parameters used as the initial value estimation model M2 may be the same as known programs and parameters. For example, the parameters may be weights and biases.

[0114] For example, the data storage unit 100 may store a training database in which the training data necessary for the learning of the initial value estimation model M2 is stored. The training data includes an input part that is input to the initial value estimation model M2 during learning, and an output part that is the correct answer during learning. The input part of the training data is basically in the same format as the input data input to the initial value estimation model M2 during estimation. The output part of the training data is basically in the same format as the output data output from the initial value estimation model M2 during estimation. Note that the input part of the training data may be in a format slightly different from the input data input to the initial value estimation model M2 during estimation. Similarly, the output part of the training data may be in a format slightly different from the output data output from the initial value estimation model M2 during estimation.

[0115] For example, the input part of the training data includes training effect information. The input part of the training data may also include the timing when the target person for training is urged. The input part of the training data may include training cost information. These pieces of information may be the same as the information described with reference to FIG. 5. The output part of the training data is a combination of reminders (the combination of reminders that is the correct answer during learning) corresponding to the effect information and the like that are the paired input parts. The training data may be created by the administrator of the estimation system 1 or may be created by a training data creation tool.

[0116] For example, if the number of training data is relatively small, there may be a possibility of obtaining an optimal solution. Therefore, the output part of the training data may be an optimal solution obtained from a relatively small number of training data. The optimal solution may be obtained by the meta-heuristic method described above. The optimal solution may be specified within the range of constraints for generating the training data. When the input part of the training data is input, the server 10 performs learning of the initial value estimation model M2 so that the output part of the training data is output. The server 10 performs learning of the initial value estimation model M2 by adjusting the parameters of the initial value estimation model M2 based on the training data.

[0117] Note that the algorithm for learning the initial value estimation model M2 may be a known algorithm used in the field of machine learning. For example, the server 10 causes the initial value estimation model M2 to learn the training data based on an algorithm such as the gradient descent method or the error backpropagation method. The loss function used during learning may also be a known loss function. The server 10 calculates the loss, which is the error between the output part of the training data and the output from the initial value estimation model M2 during learning, based on the loss function. The server 10 completes the learning when the loss becomes relatively small. When the learning is completed, the server 10 records the learned initial value estimation model M2 in the initial value estimation model storage unit 107.

[0118] The estimation unit 105 of Modification Example 8 acquires an initial value of a combination of reminders based on the effect information and cost information of each of a plurality of target persons and the initial value estimation model M2, and estimates a combination of reminders based on the initial value. For example, the estimation unit 105 inputs the effect information and cost information of each of a plurality of target persons into the initial value estimation model M2. The initial value estimation model M2 calculates an embedded representation of the effect information and cost information of each of a plurality of target persons, and outputs an initial value corresponding to the embedded representation. The estimation unit 105 acquires the initial value output from the initial value estimation model M2. The estimation unit 105 estimates a combination of reminders based on the initial value. Although the method for acquiring the initial value of the combination of reminders is different from that of the embodiment, other points are as described in the embodiment.

[0119] The estimation system 1 of Modification Example 8 stores an initial value estimation model M2 in which the relationship between the training effect information and cost information and the training combination is learned. The estimation system 1 acquires an initial value of a combination based on the effect information and cost information of each of a plurality of target persons and the initial value estimation model M2, and estimates a combination based on the initial value. Thereby, the estimation system 1 can quickly estimate an efficient combination of reminders as a whole.

[0120] Note that the estimation system 1 of Modification Example 8 may not include the functions described in the embodiment. The estimation system 1 may have only a configuration for performing a predetermined estimation based on the initial value estimated based on the initial value estimation model M2 stored in the initial value estimation model storage unit 107. In this case, the estimation system 1 does not include a configuration for estimating a combination of reminders based on effect information, cost information, and constraint information, and has only a configuration for performing a predetermined estimation based on the initial value estimated based on the initial value estimation model M2 stored in the initial value estimation model storage unit 107. According to such an aspect, since the accuracy of the initial value is increased, it is possible to accelerate the completion of estimation when estimating some combination. Such an aspect (for example, a configuration that solves other problems described in this paragraph without solving the problems described in the column of the problems to be solved by the invention) is also within the scope of the present disclosure.

[0121] [6-9. Other Modification Examples] For example, the above Modification Examples 1 to 8 may be combined.

[0122] For example, the functions described as being realized by the subject terminal 30 may be realized by the server 10, the person-in-charge terminal 20, or other computers. The processes described as being realized by the subject terminal 30 may be shared by a plurality of computers. The processes described as being realized by the server 10 may be realized by the subject terminal 30, the person-in-charge terminal 20, or other computers. The main functions of the estimation system 1 may be shared by a plurality of computers.

[0123] [7. Supplementary Note] For example, the estimation system may have the following configuration. (1) An effect information acquisition unit that acquires effect information regarding the effects obtained when a predetermined action is performed on each of a plurality of subjects by a predetermined means at a predetermined timing, A cost information acquisition unit that acquires cost information regarding the costs required when the action is performed on each of the plurality of subjects by the means at the timing, A constraint information acquisition unit that acquires constraint information regarding the constraints on the costs allowed for the action, Based on the effect information and the cost information of each of the plurality of subjects, and the constraint information, an estimation unit that estimates a combination of the subject on whom the action is to be performed, and at least one of the timing and the means, so that the effect can be obtained efficiently as a whole within the range of the constraints, An estimation system including the above. (2) The effect information acquisition unit acquires the effect information of each of the plurality of subjects based on the probability that the effect is obtained when the action is performed on each of the plurality of subjects by the means at the timing, and the degree of the effect obtained from the subject. The estimation system according to (1). (3) The estimation unit calculates an index related to the efficiency of the effect obtained from each of the plurality of target persons based on the effect information and the cost information of each of the plurality of target persons, and estimates the combination based on the index of each of the plurality of target persons. The estimation system according to (1) or (2). (4) The cost information acquisition unit calculates, for each of the plurality of target persons, the probability of a response when the action is performed by the means at the timing, and acquires the cost information based on the probability of each of the plurality of target persons. The estimation system according to any one of (1) to (3). (5) The cost information acquisition unit calculates, for each of the plurality of target persons, the probability that the effect is obtained when the action is performed by the means at the timing, and acquires the cost information based on the probability of each of the plurality of target persons. The estimation system according to any one of (1) to (4). (6) The estimation system further includes a future risk information acquisition unit that acquires, for each of the plurality of target persons, future risk information related to the risk that the action will be required in the future. The estimation unit estimates the combination based on the future risk information as well. The estimation system according to any one of (1) to (5). (7) The future risk information acquisition unit acquires the future risk information based on the probability that the risk occurs and the degree of the risk for each of the plurality of target persons. The estimation system according to (6). (8) Each of the plurality of target persons uses a predetermined service. The future risk information acquisition unit acquires the future risk information based on the past usage status of the service by each of the plurality of target persons. The estimation system according to (6) or (7). (9) The estimation unit estimates the combination so that the action is preferentially performed on the subject for whom the effect is relatively large. The estimation system according to any one of (1) to (8). (10) The constraint information acquisition unit acquires the constraint information in which the constraint according to the timing is defined. The estimation unit estimates the combination within the range of the constraint according to the timing. The estimation system according to any one of (1) to (9). (11) The constraint information acquisition unit acquires the constraint information in which the constraint according to the means is defined. The estimation unit estimates the combination within the range of the constraint according to the means. The estimation system according to any one of (1) to (10). (12) The estimation system further includes an initial value estimation model storage unit that stores an initial value estimation model in which the relationship between the training effect information and cost information, and the training combination, is learned. The estimation unit acquires an initial value of the combination based on the effect information and cost information of each of the plurality of subjects, and the initial value estimation model, and estimates the combination based on the initial value. The estimation system according to any one of (1) to (11). (13) The estimation system further includes a probability estimation model storage unit that stores a probability estimation model in which the relationship between the subject characteristic information regarding the characteristics of the subject for training, and the probability of obtaining the effect, is learned. The effect information acquisition unit acquires the probability of obtaining the effect from the subject based on the effect information of each of the plurality of subjects and the probability estimation model, and acquires the effect information based on the probability of each of the plurality of subjects. The estimation system according to any one of (1) to (12). (14) Each of the plurality of target persons is a delinquent who has defaulted on the payment of a credit card. The means is a communication means for reminding the delinquent. The action is the reminder. The cost is the cost for the reminder. The estimation system according to any one of (1) to (13).

Explanation of Signs

[0124] 1 Estimation system, 10 Server, 11, 21, 31 Control unit, 12, 22, 32 Storage unit, 13, 23, 33 Communication unit, 24, 34 Operation unit, 25, 35 Display unit, 20 Person-in-charge terminal, 30 Target person terminal, DB Target person database, M1 Probability estimation model, M2 Initial value estimation model, N Network, SC Screen, 100 Data storage unit, 101 Probability estimation model storage unit, 102 Effect information acquisition unit, 103 Cost information acquisition unit, 104 Constraint information acquisition unit, 105 Estimation unit, 106 Future risk information acquisition unit, 107 Initial value estimation model storage unit.

Claims

1. An effect information acquisition unit that acquires effect information regarding an effect obtained when a predetermined action is performed on each of a plurality of target persons by a predetermined means at a predetermined timing; A cost information acquisition unit that acquires cost information regarding the cost required when the action is performed on each of the plurality of target persons by the means at the timing; A constraint information acquisition unit that acquires constraint information regarding a constraint on the cost allowed for the action; Based on the effect information and the cost information of each of the plurality of target persons, for each combination candidate of the target person on whom the action is to be performed, at least one of the timing and the means, an index regarding the efficiency of the effect obtained from the target person is calculated, and the total value of the cost of the combination is within the range of the constraint indicated by the constraint information, and the total value of the index of the combination is relatively higher than that of the other combinations, an estimation unit that estimates the combination; An estimation system including the above.

2. The effect information acquisition unit acquires the effect information of each of the plurality of target persons based on the probability that the effect is obtained when the action is performed on each of the plurality of target persons by the means at the timing and the degree of the effect obtained from the target person. The estimation system according to claim 1.

3. The cost information acquisition unit calculates the probability of response when the action is performed on each of the plurality of target persons by the means at the timing, and acquires the cost information based on the probability of each of the plurality of target persons. The estimation system according to claim 1 or 2.

4. The cost information acquisition unit calculates the probability that the effect is obtained when the action is performed on each of the plurality of target persons by the means at the timing, and acquires the cost information based on the probability of each of the plurality of target persons. The estimation system according to claim 1 or 2.

5. The estimation system further includes a future risk information acquisition unit that acquires future risk information regarding the risk that the action will be required in the future for each of the plurality of target persons, The estimation unit calculates the index for each candidate based on a weight coefficient according to the future risk information. The estimation system according to claim 1 or 2.

6. The future risk information acquisition unit acquires the future risk information based on the probability of the risk occurring for each of the plurality of target persons and the degree of the risk. The estimation system according to claim 5.

7. Each of the plurality of target persons uses a predetermined service. The future risk information acquisition unit acquires the future risk information based on the past usage status of the service by each of the plurality of target persons. The estimation system according to claim 5.

8. The estimation unit calculates the index for each candidate based on a weight coefficient determined such that the action is preferentially performed for the target person for whom the effect is relatively large. The estimation system according to claim 1 or 2.

9. The constraint information acquisition unit acquires the constraint information in which the constraint corresponding to the timing is defined. The estimation unit estimates the combination such that the total value becomes relatively higher than the other combinations within the range of the constraint corresponding to the timing. The estimation system according to claim 1 or 2.

10. The constraint information acquisition unit acquires the constraint information in which the constraint corresponding to the means is defined. The estimation unit estimates the combination such that the total value becomes relatively higher than the other combinations within the range of the constraint corresponding to the means. The estimation system according to claim 1 or 2.

11. The estimation system further includes an initial value estimation model storage unit that stores an initial value estimation model in which the relationship between the training effect information and cost information and the training combination is learned. The estimation unit acquires an initial value of the combination based on the effect information and cost information of each of the plurality of target persons and the initial value estimation model, and changes the initial value to estimate the combination. The estimation system according to claim 1 or 2.

12. The estimation system further includes a probability estimation model storage unit that stores a probability estimation model in which the relationship between the target person characteristic information regarding the characteristics of the target person for training and the probability of obtaining the effect is learned. The effect information acquisition unit acquires the probability of obtaining the effect from the target person based on the target person characteristic information regarding the characteristics of each of the plurality of target persons and the probability estimation model, and acquires the effect information based on the probability of each of the plurality of target persons and the degree of the effect obtained from the target person. The estimation system according to claim 1 or 2.

13. Each of the plurality of target persons is a delinquent who has defaulted on credit card payments, The means is a communication means for urging the delinquent, The action is the urging, The cost is the cost for the urging, The estimation system according to claim 1 or 2.

14. A computer performs an effect information acquisition step of acquiring effect information regarding the effect obtained when a predetermined action is performed on each of the plurality of target persons by a predetermined means at a predetermined timing; a cost information acquisition step of acquiring cost information regarding the cost required when the action is performed on each of the plurality of target persons by the means at the timing; a constraint information acquisition step of acquiring constraint information regarding the constraint on the cost allowed for the action; Based on the effect information and the cost information of each of the plurality of target persons, for each combination candidate of the target person on whom the action is to be performed, at least one of the timing and the means, an index regarding the efficiency of the effect obtained from the target person is calculated, and the total value of the cost of the combination is within the range of the constraint indicated by the constraint information, and the total value of the index of the combination is relatively higher than that of the other combinations, an estimation step of estimating the combination; An estimation method for executing.

15. An effect information acquisition unit that acquires effect information regarding the effect obtained when a predetermined action is performed on each of the plurality of target persons by a predetermined means at a predetermined timing; a cost information acquisition unit that acquires cost information regarding the cost required when the action is performed on each of the plurality of target persons by the means at the timing; a constraint information acquisition unit that acquires constraint information regarding the constraint on the cost allowed for the action; Based on the effect information and the cost information of each of the plurality of target persons, for each combination candidate of the target person on whom the action is to be performed, at least one of the timing and the means, an index regarding the efficiency of the effect obtained from the target person is calculated, and the total value of the cost of the combination is within the range of the constraint indicated by the constraint information, and the total value of the index of the combination is relatively higher than that of the other combinations, an estimation unit that estimates the combination; A program for causing a computer to function as such.

Images