JP6194434B1 - Information processing apparatus, information processing method, and program - Google Patents

Abstract

An object of the present invention is to reduce a departure rate while suppressing occurrence of default. Receiving means for receiving specific information that is information that can identify an individual, acquiring means for acquiring personal credit information of the individual based on the specific information received by the receiving means, and the acquiring means Based on the personal credit information acquired by the input data determination means for determining the input data of a predetermined model, the input data determined by the input data determination means, and based on the model, Default probability determining means for determining a default probability. [Selection] Figure 9

Description

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

  Patent Document 1 describes the following loan credit judgment system as a technique related to credit judgment. The loan credit judgment system disclosed in Patent Document 1 selects factors such as attribute information, real estate information, and fund information for attribute classification based on customer information, sets a regression coefficient for each factor, and calculates an appropriate rate. When the customer presentation rate exceeds the calculated appropriate rate, the lending is approved. Further, in the loan credit judgment system of Patent Document 1, after lending, the permission judgment standard is corrected in consideration of the repayment status and the like, and based on recent information.

JP 2008-90689 A

  In the technique of Patent Document 1, attribute information, real estate information, fund information, and the like are required. In order for a financial institution to acquire such information, it is necessary for a customer to fill in various information on a loan application form. If attribute information or the like is used, the accuracy of predicting a default is improved, so that the default risk can be reduced as a financial institution. However, it is troublesome to fill in various information on the loan application form. Loan application forms also have items with a psychological burden such as annual income. For this reason, there is a tendency that the withdrawal rate at which the customer quits the procedure by interrupting the application for the loan or the like becomes high.

  An object of the present invention is to reduce the separation rate while suppressing the occurrence of default.

An information processing apparatus according to the present invention includes: a receiving unit that receives specific information that is information that can specify an individual; and an acquiring unit that acquires personal credit information of the individual based on the specific information received by the receiving unit. , Input data determination means for determining input data of a predetermined model based on the personal credit information acquired by the acquisition means, the input data determined by the input data determination means, and the model based on, possess a default probability determining means for determining a default probability, the, the input data determination unit, based on the personal credit information acquired by the acquisition means, as the input data of the model, the Determine one or more credit-related variables that are related to the credit card with which the individual has a contract, and Shopping related variables include any of shopping related variables, cashing related variables, credit usage period, total remaining credit payment, and maximum remaining credit payment, and the shopping related variables include the credit contracted by the individual. These are variables related to card shopping loans, the cashing-related variables are variables related to the credit card cashing contracted by the individual, and the credit usage period is calculated for each credit card contracted by the individual. The credit remaining payment total value is a total value of “residual bond amount / deposit amount” calculated for each credit card contracted by the individual, and the credit remaining payment maximum value. Is calculated for each credit card contracted by the individual. It has a maximum value of "Zansaigaku / deposit amount".
The information processing apparatus according to the present invention includes a receiving unit that receives specific information that is information that can identify an individual, and an acquisition that acquires personal credit information of the individual based on the specific information received by the receiving unit. Means, input data determination means for determining input data of a predetermined model based on the personal credit information acquired by the acquisition means, the input data determined by the input data determination means, and Default probability determining means for determining a default probability based on the model, the input data determining means based on the personal credit information acquired by the acquiring means as input data of the model, One or two or more borrowing history related variables that are variables related to the borrowing history are determined, and the borrowing balance is determined as the borrowing history related variable. Information, maximum balance progress information, borrowing balance increase, borrowing number increase number, borrowing balance decrease number, borrowing balance change information includes the borrowing balance at the predetermined time and the borrowing before the predetermined time The maximum balance progress information is a variable based on a maximum balance elapsed time that is an elapsed time from when the borrowing balance included in the personal credit information is maximum, and the increase in the borrowing balance is a variable based on a difference from the balance The amount is an increase in the borrowing balance within the range of the personal credit information, the increase in the number of borrowing is the increase in the number of the latest predetermined claims, and the decrease in the borrowing balance is the personal credit information This is the number of times the borrowing balance within the range decreased.
The information processing apparatus according to the present invention includes a receiving unit that receives specific information that is information that can identify an individual, and an acquisition that acquires personal credit information of the individual based on the specific information received by the receiving unit. Means, input data determination means for determining input data of a predetermined model based on the personal credit information acquired by the acquisition means, the input data determined by the input data determination means, and Default probability determination means for determining a default probability based on the model, the input data determination means as the input data of the model based on the personal credit information acquired by the acquisition means Determining at least one of the total borrowing balance and the number of exclusive contracts, the total borrowing balance is a total of the borrowing balance, Serial full-time contract number, of the contract of borrowing that the individual is doing, the personal is the number of contracts that are carried out with the loan specialist's.
The information processing apparatus according to the present invention includes a receiving unit that receives specific information that is information that can identify an individual, and an acquisition that acquires personal credit information of the individual based on the specific information received by the receiving unit. Means, input data determination means for determining input data of a predetermined model based on the personal credit information acquired by the acquisition means, the input data determined by the input data determination means, and Default probability determining means for determining a default probability based on the model, the input data determining means based on the personal credit information acquired by the acquiring means as input data of the model, Determine one or two or more arrears-related variables that are variables related to the arrears of deposit, Includes the number of months past due, the maximum number of days past due, the maximum number of days past due, the number of past due loans, and the number of past due loans. The maximum number of delinquencies, the number of consecutive delinquent months is the maximum number of consecutive delinquent months that occurred in the range of a predetermined period in the past, and the maximum number of past due days is the past predetermined number of deposits The maximum number of days in arrears for each receivable and each time within the range of the above is the maximum value, and the maximum number of days in arrears for each month is the number of days in arrears for each receivable and for each month within the range of the past specified period. The number of past due loans by number is the number of loans that have been delinquent within the range of past deposits, and the number of past due loans by month is past due within the past prescribed period. This is the number of receivables.

  According to the present invention, it is possible to reduce the departure rate while suppressing the occurrence of default.

FIG. 1 is a diagram illustrating an example of a hardware configuration of a credit system. FIG. 2 is a diagram illustrating an example of a software configuration of the credit server. FIG. 3 is a diagram illustrating an example of a screen displayed on the customer terminal. FIG. 4 is a diagram illustrating an example of a processing sequence of the credit system. FIG. 5 is a diagram illustrating an example of personal credit information. FIG. 6 is a diagram illustrating an example of the conversion table. FIG. 7 is a diagram for explaining the importance of explanatory variables. FIG. 8 is a diagram illustrating an example of an AR for each model. FIG. 9 is a diagram illustrating an example of a flowchart of credit processing. FIG. 10 is a diagram showing a CAP curve. FIG. 11 is a diagram for explaining the definition of AR. FIG. 12 is a diagram illustrating an image of a model expressed by a decision tree. FIG. 13 is a diagram illustrating an example of the second credit information.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[Overall configuration of credit system]
First, the overall configuration of the credit system 1 will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of the configuration of the credit system 1. The credit system 1 is a system that makes a credit decision such as a loan based on input data from a customer, and includes a credit server 100, a customer terminal 200, and a personal credit information server 300.

The credit server 100 is an information processing apparatus that performs credit judgment such as a loan based on input data from a customer.
The customer terminal 200 is an information processing apparatus that is used by a customer to apply for a loan or the like. As the customer terminal 200, a mobile phone including a smartphone, a tablet terminal, a personal computer, or the like is used, but is not limited thereto.
The personal credit information server 300 is an information processing apparatus that manages personal credit information, and transmits personal credit information corresponding to the request to the external information processing apparatus in response to a request from the external information processing apparatus.
The credit server 100, the customer terminal 200, and the personal credit information server 300 can communicate data via the network 10. The network 10 is the Internet or the like.

[Hardware configuration]
Next, with reference to FIG. 1, the hardware configuration of the information processing apparatus constituting the credit system 1 will be described.
First, the hardware configuration of the credit server 100 will be described. The credit server 100 includes, as hardware, a CPU 101, a storage device 102, a communication interface 103, and a bus 104 that connects them.
The CPU 101 controls the entire credit server 100. When the CPU 101 executes processing based on a program stored in the storage device 102 or the like, the software configuration of the credit server 100 and the processing of the credit server 100 shown in FIGS. 4 and 9 described later are realized.
The storage device 102 is a storage device such as a RAM, a ROM, and an HDD, and stores a program, and stores data used when the CPU 101 executes processing based on the program. The storage device 102 stores personal credit information that the credit server 100 acquires from the personal credit information server 300.
The communication interface 103 controls communication between the credit server 100 and external devices such as the customer terminal 200 and the personal credit information server 300.

Next, the hardware configuration of the customer terminal 200 will be described. The customer terminal 200 includes, as hardware, a CPU 201, a storage device 202, a communication interface 203, an input device 204, a display device 205, and a bus 206 that connects them.
The CPU 201 controls the customer terminal 200 as a whole. When the CPU 201 executes processing based on a program stored in the storage device 202 or the like, processing of the customer terminal 200 shown in FIG. 4 described later is realized.

The storage device 202 is a storage device such as a RAM, a ROM, and an HDD, and stores a program, and stores data used when the CPU 201 executes processing based on the program.
The communication interface 203 controls communication between the customer terminal 200 and an external device such as the credit server 100.
The input device 204 is a touch panel, a mouse, a keyboard, or the like, and receives a user instruction.
The display device 205 is a liquid crystal monitor or the like and can display an image.

Next, the hardware configuration of the personal credit information server 300 will be described. The personal credit information server 300 includes, as hardware, a CPU 301, a storage device 302, a communication interface 303, and a bus 304 that connects them.
Since the CPU 301, the storage device 302, and the communication interface 303 are the same as those of the credit server 100 having the same names, the description thereof is omitted. When the CPU 301 executes processing based on a program stored in the storage device 302 or the like, the processing of the personal credit information server 300 shown in FIG. 4 described later is realized.

Software configuration
Next, a software configuration for realizing the function of the credit server 100 will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of the software configuration of the credit server 100. The credit server 100 includes a reception unit 110, an acquisition unit 111, an input data determination unit 112, a default probability determination unit 113, a first screening unit 114, and a second screening unit 115.

The receiving unit 110 receives specific information from the customer terminal 200. The specific information is information that can identify an individual, and includes a name, a date of birth, and a telephone number. For the name, the name of Kana expressed in Kana is used. In the present embodiment, the specific information further includes predetermined personal identification information. The personal identification information is information used for collation with the personal identification information, and an address is used in this embodiment. The personal identification information is information that can prove the personal identification information, and for example, a driver's license or passport image data is used. In addition, as information for specifying an individual, for example, a highly reliable predetermined SNS account information or biometric information (fingerprint information, etc.) is used in accordance with the authentication method employed in the credit system. Also good.
The acquiring unit 111 acquires personal credit information from the personal credit information server 300 based on the specific information received by the receiving unit 110. The acquisition unit 111 according to the present embodiment transmits the name, date of birth, and telephone number received by the receiving unit 110 to the personal credit information server 300 that manages personal credit information. Personal credit information is acquired by receiving credit information.
Based on the personal credit information acquired by the acquisition unit 111, the input data determination unit 112 determines input data of a personal credit information detail model described later.
The default probability determining unit 113 determines a default probability based on the input data determined by the input data determining unit 112 and the personal credit information detail model.
The first screening unit 114 performs the first screening based on the default probability determined by the default probability determination unit 113. The first examination is a provisional examination of whether or not a loan is accepted.
When approved by the first examination unit 114, the second examination unit 115 performs the second examination based on the personal identification information and the personal identification information received by the reception unit 110. The second examination is a main examination of whether or not the loan is accepted. The second examination unit 115 of the present embodiment uses an address as personal identification information.

[Application screen]
Next, the application screen 400 will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of the application screen 400. The application screen 400 is a screen displayed on the display device 205 of the customer terminal 200 and is used by a customer to apply for a loan or the like using the credit system 1. The customer terminal 200 receives the specific information by the customer input on the application screen 400.
The application screen 400 includes a Kana name input unit 401 for inputting a Kana name, a birth date input unit 402 for inputting a date of birth, a telephone number input unit 403 for inputting a telephone number, an address Address input unit 404, application button 405 for executing application, and cancel button 406 for canceling application.
The CPU 201 of the customer terminal 200 may display an application screen 400 by executing an application stored in the storage device 202. Further, the CPU 201 of the customer terminal 200 may execute a WEB browser, acquire data on the application screen 400 from the credit server 100 or the like, and display the application screen 400 on the WEB browser.

[Processing sequence]
Next, the processing sequence of the credit system 1 will be described with reference to FIG. FIG. 4 is a diagram illustrating an example of a processing sequence of the credit system 1.
It is assumed that the customer terminal 200 displays the application screen 400 in FIG. 3 in advance.
In SQ100, customer terminal 200 accepts the specific information input on application screen 400. More specifically, the customer terminal 200 accepts a name, date of birth, phone number, and address.
In SQ101, when the application button 405 on the application screen 400 is pressed, the customer terminal 200 transmits the specific information received in SQ100 to the credit server 100. The credit server 100 receives the specific information.
In SQ102, the credit server 100 transmits the name, date of birth, and telephone number to the personal credit information server 300. The credit server 100 receives the name, date of birth, and telephone number.

In SQ103, the personal credit information server 300 transmits the personal credit information corresponding to the name, date of birth, and telephone number received in SQ102 to the credit server 100. The credit server 100 receives personal credit information.
In SQ104, the credit server 100 performs provisional examination. In this sequence, the credit server 100 assumes that the result of the provisional examination is provisional consent.
In SQ 105, the credit server 100 transmits the result of provisional examination (temporary consent) to the customer terminal 200. The customer terminal 200 receives the result of the provisional examination.
In SQ106, the customer terminal 200 displays the result of provisional examination. In addition, the customer terminal 200 receives the personal identification information. The personal identification information is information that can verify the personal identification information. As the personal identification information, for example, image data of a driver's license or image data of a passport can be used. When the customer terminal 200 includes a camera, the customer terminal 200 accepts the personal identification information by photographing the driver's license or the like with the camera of the customer terminal 200 by the customer's operation and acquiring the personal identification information. The customer terminal 200 may accept the personal identification information by another method such as receiving personal identification information from another device.

In SQ107, the customer terminal 200 transmits the personal identification information received in SQ106 to the credit server 100. The credit server 100 receives the personal identification information.
In SQ108, the credit server 100 performs the main examination. In this sequence, the credit server 100 assumes that the result of this examination is this consent.
In SQ 109, the credit server 100 transmits the result of the main examination (final consent) to the customer terminal 200. The customer terminal 200 receives and displays the result of this examination.

[Personal credit information]
Next, the personal credit information managed by the personal credit information server 300 will be described with reference to FIG. FIG. 5 is a diagram illustrating an example of personal credit information. The personal credit information is information representing the contract content such as a loan and the payment status.
The personal credit information includes principal information 500, borrowing information 510, credit information 520, inquiry history 530, and borrowing balance history 540.
The principal information 500 is information relating to the principal, and includes information such as name, date of birth, gender, address, telephone number, and work place. The name of the personal information 500 includes a name of Kana expressed in kana.

The borrowing information 510 is information relating to a loan receivable and is information corresponding to the Money Lending Business Law in Japan. The borrowing information 510 can include a plurality of records. Each record of the borrowing information 510 includes a company code, a balance, a payment status, and the like. One record of the borrowing information 510 represents one bond.
The company code of the borrowing information 510 is a symbol that can identify the company of the individual contracting party specified by the principal information 500. The first character of the company code represents the company type. In particular, when the first character of the company code is “S”, it indicates that the company corresponding to this company code is a loan specialist.
The balance of the borrowing information 510 is the balance of the amount borrowed from the company identified by the company code of the borrowing information 510 by the individual specified by the principal information 500.
The deposit status of the borrowing information 510 represents a past predetermined deposit status. The number of deposits is the number of days past due. For example, when “previous” is “0”, it indicates that there was no arrears in the previous deposit. Further, when “3 times before” is “10”, it means that 10 days of delinquency occurred due to the payment before 3 times.

Credit information 520 is receivable information other than loan receivables such as information related to credit, and is information corresponding to the installment sales method in Japan. Credit information 520 can include a plurality of records. The information on credit includes information on shopping transactions using a credit card. One record of the credit information 520 represents one credit card with which the customer is contracted. Each record of the credit information 520 includes a contract end date, a contract amount, a guarantee amount, a balance, and the like.
The contract end date of the credit information 520 is a completion date when the repayment is completed or a contract cancellation date.
The contract amount of the credit information 520 is the contract amount in the case of an individual contract, and the limit amount in the case of a comprehensive contract. The limit is the sum of cashing and shopping.
The guaranteed amount of the credit information 520 is a guaranteed amount at the time of contract.
The balance of the credit information 520 is the sum of the installment balance and the cashing balance. The balance of the credit information 520 may be an installment balance.

The inquiry history 530 is a history of access to personal credit information (inquiry of personal credit information), and can include a plurality of records. One record of the inquiry history 530 represents one access to personal credit information. The inquiry history 530 includes an inquiry date and time, a transaction form, and the like.
The inquiry date of the credit information 520 is the date and time when access to the personal credit information was performed.
The transaction format of the credit information 520 is a transaction format between the company that accesses the personal credit information and the individual specified by the principal information 500.

The borrowing balance history 540 is a history of borrowing balance, and the borrowing balance history 540 includes a normal history 550 and a maximum borrowing balance history 560.
The normal history 550 is a history of the borrowing balance every month from the most recent month to the month 10 months before the most recent month. The history of each month includes the number of borrowings subject to total volume regulation and the borrowing balance on the base date of the month, and the number of borrowings not subject to total volume regulation and the borrowing balance. The number of borrowings represents the number of loans. The reference date is one day or the last day of the month. The most recent month is the newest month in the normal history 550.
The maximum borrowing balance history 560 is a history of the borrowing balance on the reference date of the month in which the borrowing balance was within the range managed by the personal credit information server 300. The maximum borrowing balance history 560 is data having the same format as the data of each month of the borrowing balance history 540, and the maximum borrowing balance history 560 records the month in which the borrowing balance is maximum in the year / month format.

[Individual credit information detail model]
Next, the personal credit information detail model will be described. The personal credit information detail model is an example of a predetermined model of the present invention, and is a model in which data based on information included in the personal credit information is input data and default probability is output data. The personal credit information detail model can be obtained by performing statistical processing or the like on a plurality of personal credit information accumulated in advance. The personal credit information detail model of the present embodiment is constructed using logistic regression analysis, and is expressed by the following formulas 1 and 2.
Z = α + β1 × x1 + β2 × x2 +... + Β17 × x17 (Formula 1)
PD = 1 / (1 + exp (−Z)) (Formula 2)
In Equation 1, Z is a Z score, α is an intercept, x1 to x17 are explanatory variables, and β1 to β17 are parameters determined in advance by logistic regression analysis. PD in Equation 2 is a default probability, and exp represents an exponential function with e (base of natural logarithm) as the base.
The explanatory variables x1 to x17 are the input data of the personal credit information detail model, and the default probability PD is the output data of the personal credit information detail model. The default probability PD is the default probability of the customer represented by the personal credit information that is the basis of the explanatory variables x1 to x17. The default probability is the probability that the repayment from the borrower will be delayed under the originally contracted conditions such as the default or delinquency of the borrower, the change of the debtor category or the condition change, that is, the borrower will fall into a default state in the future It is a probability.

The intercept α and the parameters β1 to β17 are determined by the credit server 100 performing logistic regression analysis using basic information that is personal credit information accumulated in advance by the credit server 100, for example. The information processing device other than the credit server 100 may determine the intercept α and the parameters β1 to β17.
The basic information may be personal credit information of an individual credited by a financial institution or the like.
The basic information may include personal credit information of individuals not credited by the financial institution or the like in personal credit information of individuals credited by the financial institution or the like. In this case, an individual who has not been credited by a financial institution or the like is treated as if a default has occurred.

Next, explanation variables that are input data will be described. Here, explanation variables with an importance of 5% or more, which will be described later, will be described.
The explanatory variable x1 is the age of the customer. The input data determination unit 112 determines the age by calculating the age from the date of birth included in the personal information 500 of FIG. The customer is an individual represented by personal credit information.
The explanatory variable x2 is a query related variable. The inquiry related variable is a variable related to an inquiry about the customer's personal credit information. The inquiry related variables include those relating to the number of inquiries for the customer's personal credit information and those relating to the number of days. Note that the number of inquiries is counted as “1” even if there are two or more inquiries on the same day.

In the present embodiment, a part of the number of inquiries (hereinafter referred to as the number of specific inquiries) out of the number of inquiries about the customer's personal credit information is used as the inquiry related variable. The number of specific inquiries in the present embodiment is the number excluding the number of mortgage inquiries out of the number of inquiries for customer personal credit information. The mortgage inquiry is an inquiry made by a company whose mortgage is in the form of a transaction with a customer for a survey on mortgages. The number of specific inquiries can be determined from the inquiry history 530 in FIG. The input data determining unit 112 determines the number of records excluding records whose transaction form is mortgage out of the records included in the inquiry history 530 of FIG. 5 as the specific inquiry number.
Note that the query-related variables are not limited to the examples described above, and variables as shown in the following table are assumed. The count of the number of cases and the number of days may be targeted for all past histories, or may be targeted only for a predetermined period (that is, the number of queries and the number of days for a predetermined inquiry), and is not particularly limited. The input data determination unit 112 determines a query-related variable based on the query history 530 in FIG.

The explanatory variable x3 is a shopping related variable. The shopping related variable is a variable related to the loan by the customer shopping. In the present embodiment, the shopping frame consumption rate is used as the shopping related variable. Of the shopping-related variables, the shopping frame consumption rate is the consumption rate of the credit card shopping frame, and can be determined from the credit information 520 in FIG. The shopping frame digestion rate is expressed by the following formula 3.
(Shopping frame consumption rate) = (Total used balance of credit card shopping frame) / (Total credit card shopping frame) (Equation 3)

The input data determination unit 112 determines the shopping frame digestion rate according to the following procedure.
In the first step, the input data determination unit 112 extracts a credit card whose contract has not ended from the credit cards represented by the credit information 520. A credit card whose contract has not ended is a credit card without a contract end date.
In the second step, the input data determination unit 112 determines a shopping frame for each credit card extracted in the first step. The shopping frame is a credit card shopping limit, and is the larger of the contract amount and the guaranteed amount.
In the third step, the input data determination unit 112 calculates the sum of the credit card shopping frames extracted in the first step, and sets it as “total credit card shopping frame” in Equation 3.
In the fourth step, the input data determination unit 112 calculates the sum of the credit card balances extracted in the first step, and sets the sum of the used balances of the credit card shopping frames in Expression 3. The used balance of the shopping frame is a balance of shopping payment obligations by using a credit card.
In the fifth step, the input data determination unit 112 calculates the “total credit card shopping frame” and “total used balance of the credit card shopping frame” calculated in the third and fourth steps, and Equation 3: The shopping frame consumption rate is determined by calculating the shopping frame consumption rate.

  In addition, as a shopping related variable, it is not limited to the example mentioned above, The variable as shown in the following table is assumed.

  In the above table, a credit card with a shopping frame is a credit card whose shopping frame is not 0 among credit cards whose contracts have not ended. The credit card using the shopping frame is a credit card whose balance of the credit information 520 is not 0 among credit cards whose contracts are not terminated. The credit card used for the calculation of shopping related variables is the credit card with which the customer is contracted. In the above table, an example is given in which the value of each variable is calculated for all credit cards, etc., but the value of each variable is set for a predetermined credit card (for example, only the credit card in use). A configuration for calculating may be used and is not particularly limited. The variables illustrated in the above table evaluate the degree of shopping usage calculated based on the contract status of each credit card, the shopping credit for each credit card, the balance of the shopping credit, etc. in the personal credit information. This is a possible index value.

The explanatory variable x4 is a borrowing history related variable. The borrowing history related variable is a variable related to the borrowing history. The explanatory variable x4 of the present embodiment uses borrowing balance change information that is a borrowing history related variable. The borrowing balance change information is information based on the difference between the borrowing balance at a predetermined time and the borrowing balance before the predetermined time, and can be determined from the borrowing balance history 540 in FIG. The borrowing balance change information of the present embodiment is defined by the following equation 4, for example.
(Borrowing balance change information) = ((the borrowing balance subject to the total amount regulation in the most recent month) − (the borrowing balance subject to the gross amount regulation in the last ten months)) / 10 (Formula 4)
The borrowing balance subject to the total amount regulation in the most recent month and the borrowing balance subject to the total amount regulation ten months before the most recent month are information included in the normal history 550. The input data determination unit 112 determines the borrowing balance change information by calculating the borrowing balance change information based on the normal history 550 of the borrowing balance history 540 and Expression 4. The balance change information is an example, and is not limited to using the borrowing balance subject to the total amount regulation for the last 10 months.

The explanatory variable x5 is a borrowing history related variable. As the explanatory variable x5 of the present embodiment, maximum balance progress information that is a borrowing history related variable is used. The maximum balance progress information is information based on the elapsed time from when the borrowing balance is maximum, and can be determined from the borrowing balance history 540.
Here, with reference to FIG. 6, a conversion table used for determining the maximum balance progress information will be described. FIG. 6 is a diagram illustrating an example of the conversion table. The conversion table associates the range of the maximum balance elapsed time with the maximum balance elapsed time log odds. The conversion table is stored in the storage device 102 of the credit server 100. The maximum balance elapsed time is the number of days from the first day of the most recent month in the normal history 550 of FIG. 5 to the first day of the month recorded in the maximum borrowed balance history 560. Represent. The maximum balance elapsed time log odds is the log odds of the maximum balance elapsed time default probability. Maximum balance elapsed time default probability is statistical processing based on the maximum balance elapsed time determined from a plurality of personal credit information prepared in advance and the presence or absence of default that can be determined from each of the plurality of personal credit information prepared in advance. Is the default probability determined by More specifically, the maximum balance elapsed time default probability is included in the range of the maximum balance elapsed time determined from the personal credit information prepared in advance among the plurality of personal credit information prepared in advance. Default probability determined by statistical processing or the like from personal credit information.

  For example, the credit server 100 determines the maximum balance elapsed time default probability in advance as follows based on a plurality of pieces of personal credit information accumulated in advance. The credit server 100 classifies a plurality of personal credit information stored in advance into a plurality of groups for each range of maximum balance elapsed time, and sets the number of corresponding personal credit information for each group of the range of maximum balance elapsed time. Acquired and the total number A of each group within the range of the maximum balance elapsed time. Further, the credit server 100 sets the number of personal credit information indicating that a default has occurred among the personal credit information corresponding to each of the groups in the range of the maximum balance elapsed time for each group in the range of the maximum balance elapsed time. The default number is B. Then, the credit server 100 determines “A / B” in each group of the range of the maximum balance elapsed time as the maximum balance elapsed time default probability corresponding to the range of the maximum balance elapsed time.

The input data determination unit 112 determines maximum balance progress information as follows. First, the input data determination unit 112 determines the maximum balance elapsed time based on the normal history 550 and the maximum borrowing balance history 560. Next, the input data determination unit 112 acquires the maximum balance elapsed time log odds corresponding to the determined maximum elapsed balance elapsed time from the conversion table, and determines the acquired maximum balance elapsed time log odds as the maximum balance progress information. .
Thus, the maximum balance elapsed time log odds are used as the maximum balance progress information for the explanatory variable x5.

  In addition, the maximum balance elapsed time default probability is stored in the conversion table instead of the logarithmic odds of the maximum balance elapsed time, and the conversion table groups the maximum balance elapsed time and groups the maximum balance elapsed time and the maximum balance elapsed. A time default probability may be associated. At this time, the input data determination unit 112 determines the maximum balance elapsed time default probability as the maximum balance progress information, and the maximum balance elapsed time default probability is used as the maximum balance progress information for the explanatory variable x5. Further, the input data determination unit 112 may determine the maximum balance elapsed time as the maximum balance progress information. At this time, the maximum balance elapsed time is used for the explanatory variable x5. Further, the input data determination unit 112 may determine the maximum balance elapsed month, which represents the maximum balance elapsed time in months, as the maximum balance progress information.

  Note that the borrowing history related variables are not limited to the examples described above, and variables as shown in the table below are assumed. The variables illustrated in the following table can be said to be index values that can evaluate the borrowing status calculated based on at least one of the number of borrowings for each predetermined period and the balance of the borrowings.

  The increase in borrowing balance is the increase in borrowing balance subject to total amount regulation within the range of personal credit information as shown in the above table. The input data determination unit 112 refers to the normal history 550 in FIG. 5 and subtracts the borrowing balance subject to total amount regulation 10 months before the most recent month, which is the oldest month, from the borrowing balance subject to total amount regulation in the most recent month. , Increase the borrowing balance. In the example of the normal history 550 in FIG. 5, the borrowing balance subject to total amount regulation in the most recent month is “100,000”, and the borrowing balance subject to total amount regulation 10 months before the nearest month is “0”. Therefore, the input data determination unit 112 sets the increase in the borrowing balance to “100000” (= 100000−0).

  The increase in the number of borrowings, as shown in the table above, is the latest increase in the number of loans subject to total volume control. The input data determination unit 112 refers to the normal history 550 in FIG. 5 and subtracts the number of receivables subject to total amount regulation one month before the last month from the number of receivables subject to total amount restriction in the most recent month to obtain the number of borrowings increased. . In the example of the normal history 550 in FIG. 5, the number of receivables subject to total amount regulation in the most recent month is “1”. The number of receivables subject to total volume control one month before the most recent month is “1”. Therefore, the most recent increase in the number of receivables subject to total volume regulation is “0” (= 1-1). Therefore, the input data determination unit 112 sets the increase in the number of borrowings to “0”.

  As shown in the table above, the number of borrowing balance reductions is the number of times the borrowing balance subject to total amount regulation within the range of personal credit information has decreased. The input data determination unit 112 refers to the normal history 550 in FIG. 5 and counts the number of months in which the total amount regulated borrowing balance has decreased from the previous month to obtain the borrowing balance decrease count. In the example of the normal history 550 in FIG. 5, the amount of borrowings subject to total volume control in the most recent month, 3 months before the most recent month, 4 months before the most recent month, 5 months before the most recent month, and 6 months before the most recent month. Has decreased from the previous month. Therefore, the input data determination unit 112 sets the borrowing balance decrease count to “5”.

  The borrowing history information is considered to have a certain correlation with the default probability of the customer. Therefore, by using the borrowing history related variable as the explanatory variable, it becomes possible to determine the default probability based on the borrowing history, and to improve the accuracy of the default probability.

The explanatory variable x6 is the total balance. The total balance is the total of the borrowing balance and can be determined from the borrowing information 510. The input data determination unit 112 calculates the total balance of the records included in the borrowing information 510 and determines the total balance.
The explanatory variable x7 is the number of exclusive contracts. The number of exclusive contracts is the number of contracts that the customer has with the loan specialist among the borrowing contracts made by the customer, and can be determined from the borrowing information 510. The input data determination unit 112 determines the number of records whose company code is a loan specialist among the records included in the borrowing information 510 as the number of exclusive contracts. As described above, when the first character of the company code is “S”, the company corresponding to this company code is a loan specialist.

Next, the importance of explanatory variables will be described with reference to FIG. FIG. 7 is an example of a diagram illustrating the importance of explanatory variables. FIG. 7 shows explanatory variables, contents represented by the explanatory variables, standardization parameters, and importance.
The standardized parameter is a value obtained by standardizing the parameter of Equation 1 corresponding to the explanatory variable, and is, for example, a standard partial regression coefficient. The parameter corresponding to the explanatory variable xi (i is a natural number) is the parameter βi.
The importance of the explanatory variable xi is a value calculated by the following equation (5), and the larger the value, the more the explanatory variable contributes to the default probability PD.
(Importance of explanatory variable xi) = | Value obtained by standardizing parameter βi | / (| Value obtained by standardizing parameter β1 | + | Value obtained by standardizing parameter β2 | + ... + | Value obtained by standardizing parameter β17 |) (Formula 5)
As can be seen from FIG. 7, the explanatory variables x1 to x7 are explanatory variables having an importance of 5% or more.

[Comparison of models]
Next, referring to FIG. 8, the accuracy of the personal credit information detail model is compared with the accuracy of other models. FIG. 8 is a diagram illustrating an example of an AR (Accuracy ratio) for each model.
Here, the AR will be described with reference to FIGS. FIG. 10 is a diagram showing a CAP curve. The CAP curve is drawn by the following procedure.
(1) Rearrange the target data in descending order of PD (the order in which risk is predicted to be high).
(2) It is sequentially determined whether the rearranged destination is a normal destination or a DF destination (default destination).
(3) The horizontal axis represents the total cumulative composition ratio of “the number of cases determined divided by the total number of cases”.
At this time, a curve in which “vertical number of determined defects ÷ total number of defective cases” is plotted as a vertical axis is a defective cumulative composition ratio curve (CAP curve). In addition, a curve in which “vertical number determined / total number of normal cases” is plotted as a vertical axis is a normal cumulative composition ratio curve.
If the CAP curve is plotted with a perfect model having perfect accuracy, a straight line is drawn up to 100% on the vertical axis at the horizontal axis point of “total number of defectives / total number of cases” as shown by the broken line in FIG. After that, it is drawn right along the upper base. On the other hand, in the case of the indiscriminate model, the CAP curve is a diagonal line as shown by a solid line without plotting in FIG. The closer the CAP curve is to the line of the perfect model, the higher the discriminating power of the model.

FIG. 11 is a diagram for explaining the definition of AR. The area surrounded by the indiscriminate model CAP curve (diagonal line) and the CAP curve actually drawn by the model is A, and the area surrounded by the perfect model CAP curve and the CAP curve actually drawn by the model is B.
At this time, a value defined as A / A + B is AR (accumulation ratio). If the accuracy of the model is low, the CAP curve approaches a diagonal line, and the numerator A approaches zero. On the other hand, if the accuracy is high, the CAP curve approaches the perfect model, and the numerator A and the denominator A + B approach each other. That is, the range that AR can take is 0 ≦ AR ≦ 1, and the closer to 1, the higher the accuracy of the model.

Hereinafter, a model to be compared with the personal credit information detail model will be described. Here, an attribute model and an attribute personal credit information detail model are listed as models to be compared with the personal credit information detail model.
The attribute model is different from the personal credit information detail model in that only attribute data is used as an explanatory variable. The attribute data is data entered by the customer in a contract application form such as a loan. In the attribute model, the attribute data includes job type, number of employees, length of service, industry, housing style, working style, unmarried, age at employment, number of children, number of family members living together, age, gender, industry and scale, family status, annual income. Information about borrowing is used.
The attribute personal credit information detail model is different from the personal credit information detail model in that attribute data is used as an explanatory variable in addition to the explanatory variable of the personal credit information detail model. In the attribute personal credit information detail model, information relating to a residence form, length of service, and annual income is used as attribute data.

FIG. 8 shows the AR for construction and verification for each model. The AR of construction is the AR when the personal credit information and attribute data used for construction of the model are used. The AR of verification is an AR when using personal credit information and attribute data that were not used for model construction.
As can be seen from FIG. 8, in the personal credit information detail model, the over-learning is not performed because the constructed AR and the verified AR are approximate values. In addition, the personal credit information detail model has higher accuracy than the attribute model. Further, it can be said that the personal credit information detail model has substantially the same accuracy as the attribute personal credit information detail model. In other words, according to the personal credit information specification model, it is possible to determine credit risk with high accuracy without using various attribute information.

[Credit processing]
Next, the credit processing will be described with reference to FIG. FIG. 9 is a diagram illustrating an example of a flowchart of credit processing. The credit processing is processing performed by the credit server 100 in the processing sequence shown in FIG.
In S <b> 100, the receiving unit 110 receives specific information from the customer terminal 200. The receiving unit 110 receives the name, date of birth, telephone number, and address as specific information.
In S101, the acquisition unit 111 transmits the specific information received in S100 to the personal credit information server 300, and requests the personal credit information server 300 for the personal credit information corresponding to the specific information received in S100. To do. The acquisition unit 111 transmits the name, date of birth, and telephone number as specific information.
In S102, the acquisition unit 111 receives the personal credit information from the personal credit information server 300, and acquires the personal credit information corresponding to the specific information received in S100.

In S103, the input data determination unit 112 determines explanatory variables x1 to x17 which are input data of the personal credit information detail model based on the personal credit information received in S102. The method for determining the explanatory variables x1 to x7 is as described above.
In S104, the default probability determination unit 113 determines the default probability based on the explanatory variables x1 to x17 determined in S103 and the personal credit information detail model. More specifically, the default probability determination unit 113 calculates the Z score by substituting the explanatory variables x1 to x17 determined in S103 into Equation 1, and substitutes the calculated Z score into Equation 2 to determine the default probability. The default probability PD is determined by calculating PD.

In S105, the first examination unit 114 performs a provisional examination as to whether or not to accept a contract such as a loan based on the default probability determined in S104. More specifically, the first examination unit 114 obtains a value obtained by subtracting a default probability determined in S104 and an interest rate representing a predetermined expense from the interest rate of a contract such as a predetermined loan. When the value is greater than or equal to a predetermined value, it is determined that the provisional consent is accepted, and when the value is less than the predetermined value, it is determined that the provisional consent is not made and the rejection is made.
In S106, the first examination unit 114 determines whether or not provisional consent has been made in S105. The first examining unit 114 advances the process to S107 when provisional consent is obtained in S105, and advances the process to S108 when provisional consent is not obtained.

In S <b> 107, the first examination unit 114 transmits to the customer terminal 200 that it is provisional consent as a result of the provisional examination.
In S <b> 108, the first examination unit 114 transmits a rejection message as a result of the provisional examination to the customer terminal 200.
In S <b> 109, the second examination unit 115 receives the personal identification information from the customer terminal 200.

In S110, the second examination unit 115 performs a main examination of whether or not the loan is approved based on the personal identification information received in S109 and the address which is the personal identification information received in S100. More specifically, the second examination unit 115 extracts an address that is identification information from the identification information received in S109. When the address extracted from the personal identification information matches the address received in S100, the second examination unit 115 determines that the personal identification has succeeded and is accepted. When the address extracted from the personal identification information does not match the address received in S100, the second examination unit 115 determines that the personal identification has failed and is rejected without acceptance.
In S111, the second examination unit 115 determines whether or not the present approval is accepted in S110. The second examination unit 115 advances the process to S112 when the main approval is made in S110, and advances the process to S113 when the main approval is not accepted.
In S <b> 112, the second examination unit 115 transmits to the customer terminal 200 that the agreement is accepted as a result of the examination.
In S113, the second examination unit 115 transmits a rejection message to the customer terminal 200 as a result of the examination.

[effect]
As described above, the credit server 100 receives specific information from the customer terminal 200, acquires personal credit information based on the received specific information, and determines a default probability based on the acquired personal credit information. Therefore, the default probability can be determined with a predetermined accuracy. In addition, the default probability can be determined from the specific information, and it is possible to determine whether the contract is accepted or not. Therefore, the customer does not need to fill in various information such as a loan application form. In addition, the customer does not need to input items that have a psychological burden on entry such as annual income. For this reason, it is possible to reduce the departure rate while suppressing the occurrence of default.
Further, the credit server 100 receives the name, date of birth, and telephone number as specific information from the customer terminal 200. The credit server 100 transmits the name, date of birth, and telephone number to the personal credit information server, receives the personal credit information from the personal credit information server, and determines the default probability. Therefore, since the customer can apply for a loan or the like with only a small amount of information input, the withdrawal rate can be reduced.

Further, the credit server 100 determines input data of the personal credit information detail model based on the personal credit information. Therefore, compared to the case where the data included in the personal credit information is used as it is, the input data that contributes to the improvement of the accuracy of the default probability can be used, so that the accuracy of the default probability determined by the credit server 100 can be ensured.
In particular, the credit server 100 uses, as input data for a personal credit information specification model, the number of specific inquiries, a shopping frame consumption rate, borrowing balance change information, maximum balance progress information, total balance, and a special contract based on personal credit information. Determine the number of cases. These are highly important data as shown in FIG. 7, and the accuracy of the default probability determined by the credit server 100 can be ensured.

The number of specific inquiries is the number excluding the number of mortgage inquiries out of the number of inquiries for customer personal credit information. Therefore, the number of specific inquiries represents the number of inquiries relating to transaction forms that are considered to have a higher default occurrence rate than mortgages. Therefore, the number of specific queries contributes to determining the default probability with high accuracy.
The shopping frame consumption rate is the consumption rate of the credit card shopping space, and is considered to have a strong correlation with the default probability. Therefore, the shopping frame consumption rate contributes to determining the default probability with high accuracy.
The borrowing balance change information is information based on a difference between a borrowing balance at a predetermined time and a borrowing balance before the predetermined time, and represents a tendency of the borrowing balance to change. Such borrowing balance is considered to have a strong correlation with the default probability. Accordingly, the borrowing balance change information contributes to determining the default probability with high accuracy.

The maximum balance progress information is information based on the elapsed time from when the borrowing balance is maximum, and is considered to have a strong correlation with the default probability. Therefore, the maximum balance progress information contributes to determining the default probability with high accuracy.
The total balance is the total of the borrowing balance, and the default probability is considered to be higher as the borrowing balance is larger. That is, the total balance is considered to have a strong correlation with the default probability. Therefore, the total balance contributes to determining the default probability with high accuracy.
The number of exclusive contracts is the number of contracts that the customer has with the loan specialist among the borrowing contracts made by the customer. The default probability is considered higher as the number of exclusive contracts increases. In other words, the number of exclusive contracts is considered to have a strong correlation with the default probability. Therefore, the number of exclusive contracts contributes to determining the default probability with high accuracy.

<Other embodiments>
In the personal credit information detail model of the above embodiment, there are 17 explanatory variables as input data from x1 to x17. However, the input data is not limited to 17 pieces. For example, the total borrowing balance determined from the borrowing information 510 and the number of exclusive contracts may not be included in the input data. Even in this case, the departure rate can be reduced, and the accuracy of the default probability determined by the credit server 100 can be ensured.
The combination of explanatory variables is arbitrary, but as can be seen from the above-described explanation of the importance, when a query-related variable is included, the default prediction accuracy determined by the credit server 100 increases. Moreover, when the shopping related variable is included in addition to the inquiry related variable, the accuracy of the default probability determined by the credit server 100 is further increased.
It should be noted that other types of explanatory variables may be added without limitation.

  The personal credit information managed by the personal credit information server 300 may include second credit information 570 shown in FIG. The second credit information 570 is information related to a credit card with which the customer is contracted. One record of the credit information 520 represents one credit card with which the customer is contracted. Each record of the second credit information 520 includes an inquiry date, a contract date, a cashing frame, a used balance of the cashing frame, a remaining bond amount, a deposit amount, and the like.

The inquiry date of the second credit information 520 is the date when the credit server 100 acquires the personal credit information from the personal credit information server 300.
The contract date of the second credit information 520 is the date on which the customer has contracted for the credit card.
The cashing frame of the second credit information 520 is a cashing available amount.
The used balance of the cashing frame of the second credit information 520 is the balance of the cash payment obligation due to the use of the credit card.
The amount of remaining bonds in the second credit information 520 is the total amount that the customer must pay to the credit company or the like.
The deposit amount of the second credit information 520 is an amount deposited by the customer to the credit company or the like.

  In the above embodiment, a shopping related variable is used as the explanatory variable x3. However, credit-related variables other than shopping-related variables may be used as explanatory variables. The credit related variable is a variable related to the credit card with which the customer is contracted. The credit-related variables include the shopping-related variables described above. In addition, variables shown in the table below are assumed as credit-related variables.

  The credit usage period is the maximum value of the usage period calculated for each credit card as shown in the above table. The usage period is expressed in units of years, for example. The input data determination unit 112 refers to the second credit information 570 in FIG. 13 and calculates the usage period from the reference date and the contract date for each credit card. The usage period is calculated by, for example, “year of reference date−year of contract date”. Next, the input data determination unit 112 sets the calculated maximum usage period as the credit usage period. In the example of FIG. 13, the maximum value is eight years of the third card usage period. Therefore, the input data determination unit 112 sets the credit usage period to “8”.

  As shown in the above table, the total credit remaining payment value is the total value of the “residual amount / deposit amount” calculated for each credit card. Further, the maximum credit remaining payment is the maximum value of the “residual bond amount / deposit amount” calculated for each credit card as shown in the above table. The input data determination unit 112 can determine the total remaining credit payment value and the maximum remaining credit payment value based on the second credit information 570 of FIG.

  As shown in the above table, the cashing-related variables are variables related to the credit card cashing contracted by the customer. More specifically, the following variables are listed. In the table below, an example of calculating the value of each variable for all credit cards, etc. is given, but the value of each variable for a predetermined credit card (for example, only the credit card in use) is given. A configuration for calculating may be used and is not particularly limited. The variables illustrated in the table below can evaluate the degree of use of cashing calculated based on the contract status of each credit card, the cashing frame for each credit card, the balance of the cashing frame, etc. in the personal credit information It can be said that it is a good index value.

  The input data determination unit 112 can determine a caching related variable based on the second credit information 570 of FIG. Note that the credit card using the cashing frame is a credit card in which the used balance of the cashing frame of the second credit information 570 is not zero. A credit card with a caching frame is a credit card in which the caching frame of the second credit information 570 is not zero.

  Credit card information including customer credit card shopping information and cashing information is considered to have a certain correlation with the default probability of the customer. Therefore, by using a credit-related variable as an explanatory variable, it is possible to determine a default probability based on information on a customer's credit card, and it is possible to improve the accuracy of the default probability.

  In addition, delinquency related variables may be used as explanatory variables. Arrears-related variables are variables related to deposit arrears, and more specifically, the variables in the table below are listed. The variables illustrated in the table below can be said to be index values that can be used to evaluate the arrears status calculated based on the payment status for each repayment date by credit.

As shown in the above table, the number of consecutive delinquencies is the maximum value of the number of consecutive delinquencies that occurred within a predetermined range of past payments (for example, 6 times or 12 times). The input data determination unit 112 refers to the deposit status of the borrowing information 510 in FIG. 5 and acquires the number of times that deposit delay has occurred continuously. Next, the input data determination unit 112 sets the maximum value of the acquired number of times as the number of consecutive arrears.
In the borrowing information 510 of FIG. 5, within the past six deposits, the first and third receivables were arreared three times in a row three times before, twice before, and three times before. ing. In the past six deposits, no arrears occurred four times before. Therefore, the input data determination unit 112 sets the number of consecutive delinquency as 3 within the past 6 deposits.
In addition, within the past 12 deposits, the first, second and third claims have been delayed five times in a row from the seventh to the eleventh. Therefore, the input data determination unit 112 sets the number of consecutive delinquency as 5 within the past 12 deposits.

As shown in the above table, the maximum number of days in arrears is the maximum value when the number of days in arrears is tabulated by the number of receivables and by number of receivables within the past predetermined number of deposits (for example, 6 or 12). The input data determination unit 112 refers to the payment status of the borrowing information 510 in FIG. 5 and sets the maximum value when the number of days past due by credit is tabulated for each loan as the maximum number of days past due by loan.
In the borrowing information 510 in FIG. 5, the number of days in arrears is tabulated in advance for each receivable and for each transaction in the deposit status. In the borrowing information 510 of FIG. 5, the number of days past due for the third deposit 15 is the largest in the past six deposits. Therefore, the input data determination unit 112 sets the maximum number of past days in arrears to 15 within the past six deposits.
Further, in the past 12 deposits, the number of days 30 days past due for deposit of the first bond is the largest. Therefore, the input data determination unit 112 sets the maximum number of past due days 30 for the past 12 deposits.

As shown in the table above, the number of past due loans is the number of loans that have been delinquent within the past predetermined number of deposits (for example, 6 times or 12 times). The input data determination unit 112 refers to the deposit status of the borrowing information 510 in FIG. 5 and sets the number of past due loans in the past predetermined deposit range as the number of past due loans.
In the borrowing information 510 of FIG. 5, within the past six deposits, the first and third receivables are overdue. Therefore, the input data determination unit 112 sets the number of past due loans by 2 within the range of the past six deposits.
In the borrowing information 510 of FIG. 5, within the past twelve deposits, the first loan, the second bond, and the third bond have been delayed. Therefore, the input data determination unit 112 sets the number of past due loans by 3 within the past 12 deposits.

  When the depositing status of the borrowing information 510 is represented not by the times as shown in FIG. 5 but by month, the input data determining unit 112 refers to the depositing status of the borrowing information 510 represented by month, The number of consecutive arrears, the maximum number of days in arrears, and the number of monthly arrears can be determined in the same manner as the number of consecutive arrears, the maximum number of days in arrears, and the number of loans in arrears.

  The customer's past payment arrears information is important information for predicting future payment arrears. In general, examiners who make credits at financial institutions attach importance to information on payment arrears. Therefore, by using the past due related variable as the explanatory variable, it becomes possible to determine the default probability based on the past payment information of the customer, and to improve the accuracy of the default probability.

  The personal credit information detail model of the above embodiment uses logistic regression analysis. However, the personal credit information detail model may use a decision tree, a support vector machine, group learning, or a neural network.

In the logistic regression analysis model, the entire model is represented by a mathematical formula for calculating one default rate. However, the model to which the present invention can be applied has a relationship between various variables determined from personal credit information and default probabilities. Any model can be used as long as it is generalized, and the model is not limited to that expressed as a single mathematical expression as a whole.
FIG. 12 is a diagram showing an image of a model expressed by a decision tree. As shown in FIG. 12, the decision tree determines the default probability PD based on the value of each input variable and the relationship between the input variables, and is not represented by one mathematical expression as a whole. That is, when the default probability PD is determined by a decision tree, a branch condition for each variable and a variable to be determined next as a result of the branch are predetermined, and a program for determining such a conditional branch Or a database (or data file) is used.

In the above embodiment, the credit server 100 determines the age of the input data of the personal credit information detail model based on the personal credit information. However, the credit server 100 may determine the age of the input data of the personal credit information detail model based on the date of birth received from the customer terminal 200.
In the above embodiment, the credit server 100 extracts an address from the personal identification information in S110 of FIG. However, the operator of the credit system 1 may input the address described in the personal identification information into the credit server 100 with reference to the personal identification information. At this time, the credit server 100 performs the main examination in S110 of FIG. 9 using the address input by the operator.
Further, the credit server 100 may perform processing up to the provisional examination without performing the main examination. At this time, the application screen 400 of FIG. 3 may not include the address input unit 404, and the credit server 100 may not receive the address from the customer terminal 200.

  Although the present invention has been described together with the embodiments, the above-described embodiments are merely examples of implementation in carrying out the present invention, and the technical scope of the present invention is interpreted in a limited manner by these. It must not be. That is, the present invention can be implemented in various forms without departing from the technical idea or the main features thereof.

1 Credit System 100 Credit Server 200 Customer Terminal 300 Personal Credit Information Server

Claims (20)

Receiving means for receiving specific information that is information that can identify an individual;
Obtaining means for obtaining personal credit information of the individual based on the specific information received by the receiving means;
Input data determining means for determining input data of a predetermined model based on the personal credit information acquired by the acquiring means;
Default probability determining means for determining a default probability based on the input data determined by the input data determining means and the model;
I have a,
The input data determining means is based on the personal credit information acquired by the acquiring means, and the input data of the model is one or more credit-related variables that are variables related to a credit card with which the individual has a contract. Determine the variables,
The credit-related variables include any of shopping-related variables, cashing-related variables, credit usage period, credit remaining payment total value, credit remaining payment maximum value,
The shopping-related variable is a variable related to a loan by shopping of a credit card with which the individual is contracted,
The cashing related variable is a variable related to cashing of a credit card with which the individual is contracted,
The credit usage period is the maximum value of the usage period calculated for each credit card contracted by the individual,
The credit residual payment total value is a total value of “residual bond amount / deposit amount” calculated for each credit card contracted by the individual,
The maximum credit remaining payment value is an information processing apparatus that is a maximum value of “amount of remaining bonds / deposit amount” calculated for each credit card with which the individual contracts. The shopping-related variable evaluates the degree of use of shopping calculated based on at least one of the contract status of each credit card, the shopping frame for each credit card, or the balance of the shopping frame in the personal credit information. Possible index values,
The cashing-related variable can evaluate the degree of use of cashing calculated based on at least one of the contract status of each credit card, the cashing frame for each credit card, and the balance of the cashing frame in the personal credit information. 2. The information processing apparatus according to claim 1 , wherein the information value is an accurate index value. As the shopping-related variables, any of the shopping frame consumption rate, the shopping predetermined ratio over number, the total shopping balance, the shopping usage rate, the shopping holding number, the shopping usage total extreme, and the shopping usage number are included.
The shopping frame consumption rate is a value calculated by “total of used balance of shopping frame of predetermined credit card / total of shopping frame of predetermined credit card”,
The above-mentioned shopping predetermined ratio excess number is a value calculated by “the number of credit cards whose individual shopping frame consumption rate exceeds a predetermined ratio / the number of credit cards having a predetermined shopping frame”,
The individual shopping frame consumption rate is a value calculated by “used credit card shopping frame / credit card shopping frame”,
The shopping total balance is the total of the shopping credit balance of a predetermined credit card,
The shopping utilization rate is a value calculated by “the number of credit cards using a shopping frame / the number of credit cards having a predetermined shopping frame”,
The shopping holding number is the number of credit cards with a predetermined shopping frame,
The shopping use total extreme is a total of shopping frames of a predetermined credit card,
The shopping number available, the information processing apparatus according to claim 1 or 2 wherein the number of credit cards that use the shopping frame. The caching-related variables include any of the following: caching frame digestion rate, more than a predetermined percentage of caching, total caching balance, caching usage rate, number of cached holdings, total caching extreme, number of caching usage,
The cashing frame consumption rate is a value calculated by “the sum of the used balance of the cashing frame of the predetermined credit card / the total of the cashing frame of the predetermined credit card”,
The caching predetermined ratio super number is a value calculated by “the number of credit cards whose individual caching frame digestion rate exceeds a predetermined ratio”,
The individual cashing frame consumption rate is a value calculated by “used balance of credit card cashing frame / credit card cashing frame”,
The total cash balance is the sum of used balances of a predetermined credit card cash frame,
The caching usage rate is a value calculated by “the number of credit cards using a cashing frame / the number of credit cards having a predetermined cashing frame”.
The cashing possession number is the number of credit cards having a predetermined cashing frame,
The total cashing extreme is the sum of the predetermined credit card cashing frames,
The information processing apparatus according to any one of claims 1 to 3 , wherein the number of cashing uses is a number of credit cards using a cashing frame. Receiving means for receiving specific information that is information that can identify an individual;
Obtaining means for obtaining personal credit information of the individual based on the specific information received by the receiving means;
Input data determining means for determining input data of a predetermined model based on the personal credit information acquired by the acquiring means;
Default probability determining means for determining a default probability based on the input data determined by the input data determining means and the model;
Have
The input data determining means determines one or more borrowing history related variables that are variables relating to borrowing history as input data of the model based on the personal credit information acquired by the acquiring means,
The borrowing history related variable includes any of borrowing balance change information, maximum balance progress information, borrowing balance increase amount, borrowing number increase number, borrowing balance decrease number,
The borrowing balance change information is a variable based on a difference between a borrowing balance at a predetermined time and a borrowing balance before the predetermined time,
The maximum balance progress information is a variable based on a maximum balance elapsed time that is an elapsed time from when the borrowing balance included in the personal credit information is maximum,
The borrowing balance increase is an increase in the borrowing balance within the range of the personal credit information,
The increase in the number of borrowings is the increase in the number of the latest predetermined claims,
The number of times the borrowing balance is reduced is the number of times that the borrowing balance is reduced within the range of the personal credit information. The information processing apparatus according to claim 5, wherein the borrowing history related variable is an index value capable of evaluating a borrowing status calculated based on at least one of the number of borrowings for each predetermined period and the balance of the borrowings. The input data determining means determines information based on a maximum balance elapsed time default probability corresponding to the maximum balance elapsed time as the maximum balance elapsed information based on the maximum balance elapsed time and a conversion table,
The conversion table associates the range of the maximum balance elapsed time with information based on the maximum balance elapsed time default probability,
The maximum balance elapsed time default probability is included in the range of the maximum balance elapsed time determined from the personal credit information prepared in advance among the plurality of personal credit information prepared in advance. The information processing apparatus according to claim 5 or 6, wherein the default probability is determined from the personal credit information. Receiving means for receiving specific information that is information that can identify an individual;
Obtaining means for obtaining personal credit information of the individual based on the specific information received by the receiving means;
Input data determining means for determining input data of a predetermined model based on the personal credit information acquired by the acquiring means;
Default probability determining means for determining a default probability based on the input data determined by the input data determining means and the model;
Have
The input data determination means determines, as the input data of the model, at least one of the total borrowing balance and the number of exclusive contracts based on the personal credit information acquired by the acquisition means,
The total borrowing balance is the total borrowing balance,
The number of exclusive contracts is an information processing apparatus that is the number of contracts that the individual has with a loan specialist among borrowing contracts that the individual has. Receiving means for receiving specific information that is information that can identify an individual;
Obtaining means for obtaining personal credit information of the individual based on the specific information received by the receiving means;
Input data determining means for determining input data of a predetermined model based on the personal credit information acquired by the acquiring means;
Default probability determining means for determining a default probability based on the input data determined by the input data determining means and the model;
Have
The input data determination means determines one or more arrears-related variables that are variables related to payment arrears as input data of the model based on the personal credit information acquired by the acquisition means,
The arrears-related variables include any of the number of consecutive delinquencies, the number of consecutive delinquencies, the maximum number of past due days, the maximum number of past due days, the number of past due loans, the number of past due loans,
The number of consecutive delinquency is the maximum value of the number of consecutive delinquency that occurred in the past predetermined number of deposits,
The consecutive months of delinquency is the maximum number of months of continuous delinquency that occurred in the range of a predetermined period in the past,
The maximum number of days in arrears is the maximum value when the number of days in arrears is tabulated by receivable, and by the number of days in arrears, within the range of past predetermined number of deposits,
The monthly maximum arrears days are the maximum value when the arrears days are tabulated for each receivable and for each month within the range of the past predetermined period,
The number of past due loans by number is the number of past due loans within the prescribed number of deposits in the past,
The number of past due loans by month is an information processing device that is the number of past due loans in a predetermined period in the past. The information processing apparatus according to claim 9 , wherein the arrears-related variable is an index value capable of evaluating an arrears status calculated based on a payment status for each repayment date for each loan. The input data determining means is one or more inquiry related variables that are variables related to an inquiry about the personal credit information of the individual as the input data of the model based on the personal credit information acquired by the acquiring means. claims 1 to 10 any one information processing apparatus according determining variable. The information processing apparatus according to claim 11 , wherein the inquiry-related variable includes a total number of predetermined inquiries or inquiries for the individual credit information of the individual in a predetermined period. The information processing apparatus according to claim 11 , wherein the inquiry-related variable includes a number of specific inquiries excluding a predetermined number of inquiries out of the number of inquiries about the personal credit information of the individual in a predetermined period. The information processing apparatus according to claim 11 , wherein the inquiry related variable includes a specific inquiry day excluding a predetermined number of days of inquiry for the personal credit information of the individual in a predetermined period. The receiving means receives the name, date of birth, and telephone number as specific information,
The acquisition means transmits the name, date of birth, and telephone number received by the receiving means to a personal credit information server that manages the personal credit information, and the personal credit information is transmitted from the personal credit information server. by receiving information processing apparatus according to claim 1 to 14 any one of claims acquires the personal credit information. An information processing method executed by an information processing apparatus,
A receiving step of receiving specific information that is information that can identify an individual;
On the basis of the specific information received by the receiving step, an acquisition step of acquiring personal credit information of said individual,
An input data determining step for determining input data of a predetermined model based on the personal credit information acquired by the acquiring step;
A default probability determining step for determining a default probability based on the input data determined by the input data determining step and the model;
I have a,
In the input data determination step, based on the personal credit information acquired in the acquisition step, as the input data of the model, one or more credit-related variables that are variables relating to a credit card with which the individual is contracted Determine the variables,
The credit-related variables include any of shopping-related variables, cashing-related variables, credit usage period, credit remaining payment total value, credit remaining payment maximum value,
The shopping-related variable is a variable related to a loan by shopping of a credit card with which the individual is contracted,
The cashing related variable is a variable related to cashing of a credit card with which the individual is contracted,
The credit usage period is the maximum value of the usage period calculated for each credit card contracted by the individual,
The credit residual payment total value is a total value of “residual bond amount / deposit amount” calculated for each credit card contracted by the individual,
The maximum credit remaining payment value is an information processing method that is a maximum value of “amount of remaining bonds / deposit amount” calculated for each credit card contracted by the individual. An information processing method executed by an information processing apparatus,
A receiving step of receiving specific information that is information that can identify an individual;
Obtaining the personal credit information of the individual based on the specific information received by the receiving step;
An input data determining step for determining input data of a predetermined model based on the personal credit information acquired by the acquiring step;
A default probability determining step for determining a default probability based on the input data determined by the input data determining step and the model;
Have
In the input data determination step, based on the personal credit information acquired in the acquisition step, one or more borrowing history related variables that are variables relating to the borrowing history are determined as input data of the model,
The borrowing history related variable includes any of borrowing balance change information, maximum balance progress information, borrowing balance increase amount, borrowing number increase number, borrowing balance decrease number,
The borrowing balance change information is a variable based on a difference between a borrowing balance at a predetermined time and a borrowing balance before the predetermined time,
The maximum balance progress information is a variable based on a maximum balance elapsed time that is an elapsed time from when the borrowing balance included in the personal credit information is maximum,
The borrowing balance increase is an increase in the borrowing balance within the range of the personal credit information,
The increase in the number of borrowings is the increase in the number of the latest predetermined claims,
The information processing method, wherein the borrowing balance decrease count is the number of times the borrowing balance is reduced within the range of the personal credit information. An information processing method executed by an information processing apparatus,
A receiving step of receiving specific information that is information that can identify an individual;
Obtaining the personal credit information of the individual based on the specific information received by the receiving step;
An input data determining step for determining input data of a predetermined model based on the personal credit information acquired by the acquiring step;
A default probability determining step for determining a default probability based on the input data determined by the input data determining step and the model;
Have
In the input data determination step, based on the personal credit information acquired in the acquisition step, as the input data of the model, determine at least one of the total borrowing balance and the number of exclusive contracts,
The total borrowing balance is the total borrowing balance,
The information processing method, wherein the number of exclusive contracts is the number of contracts that the individual has with a loan specialist out of the borrowing contracts that the individual has performed. An information processing method executed by an information processing apparatus,
A receiving step of receiving specific information that is information that can identify an individual;
Obtaining the personal credit information of the individual based on the specific information received by the receiving step;
An input data determining step for determining input data of a predetermined model based on the personal credit information acquired by the acquiring step;
A default probability determining step for determining a default probability based on the input data determined by the input data determining step and the model;
Have
In the input data determination step, based on the personal credit information acquired in the acquisition step, as input data of the model, one or two or more arrears related variables that are variables related to payment arrears are determined;
The arrears-related variables include any of the number of consecutive delinquencies, the number of consecutive delinquencies, the maximum number of past due days, the maximum number of past due days, the number of past due loans, the number of past due loans,
The number of consecutive delinquency is the maximum value of the number of consecutive delinquency that occurred in the past predetermined number of deposits,
The consecutive months of delinquency is the maximum number of months of continuous delinquency that occurred in the range of a predetermined period in the past,
The maximum number of days in arrears is the maximum value when the number of days in arrears is tabulated by receivable, and by the number of days in arrears, within the range of past predetermined number of deposits,
The monthly maximum arrears days are the maximum value when the arrears days are tabulated for each receivable and for each month within the range of the past predetermined period,
The number of past due loans by number is the number of past due loans within the prescribed number of deposits in the past,
The number of past due loans by month is an information processing method that is the number of past due loans in a predetermined period in the past. A program for causing a computer to function as each unit of the information processing apparatus according to any one of claims 1 to 15 .

Images