KR101834007B1 - Personal credit rating device, method and computer programs based on psychometric data - Google Patents

Abstract

The present invention relates to a personal credit evaluation apparatus, method, and computer program based on psychometric measurement data, in which a credit rating evaluation standard is set using a logistic regression or machine learning technique, A first evaluating unit for evaluating the rating, a psychological evaluation criterion is generated by analyzing a correlation between an individual's psychological characteristic and a credit risk, a numerical value of psychological measurement data inputted as a response to the psychological measurement questionnaire is digitized, A second evaluation unit for evaluating the psychological class using a logistic regression model or a machine learning technique on the basis of the numerical psychometric measurement data, and a second evaluation unit for evaluating the psychological grade by the first evaluation unit, And a comprehensive evaluation unit for calculating the final credit rating by combining the psychological grades evaluated by the evaluator. According to the present invention, the credit rating of an individual can be further refined and a default can be more accurately predicted.

Description

TECHNICAL FIELD [0001] The present invention relates to personal credit evaluation devices, methods, and computer programs based on psychometric measurement data.

The present invention relates to a personal credit evaluation apparatus, method, and computer program based on psychometric measurement data, and more particularly, to a personal credit evaluation apparatus, method, and computer program based on psychometric measurement data, And a computer program.

When an individual wants to take out a loan, the credit rating is generally used as the most important evaluation criterion in loan review. Credit ratings are based on an individual's accumulated credit history. However, in the case of an individual who has a shortage of information on the accumulated credit history due to a short period of credit or a short period of credit, such a credit rating is disadvantageously applied, and the approval rate for such individuals is reduced.

In order to solve such problems, there is an effort to evaluate credit by additionally using information stored in a mobile phone, a Social Network Services (SNS) related record, and an Internet access record in addition to a credit rating. However, such information is difficult to be collected in a short time, and accumulated information is required for a certain period of time as in the case of a credit rating, so that it is difficult to apply the information immediately in an individual's credit evaluation.

In addition, in the credit evaluation, since the personal information such as the credit check history and the education level can not be used under the domestic laws and ordinances, there is a need for a method of evaluating an individual's credit on a variety of criteria in addition to the conventional credit rating.

Published Japanese Patent Application No. 10-2001-0096448 (November 11, 2001)

The present invention evaluates the credit rating based on information input from an individual who wants to receive the loan, evaluates the psychological rating based on the psychometric measurement data obtained through the psychometric questionnaire, and combines the evaluated credit rating and psychological rating A personal credit evaluation device, method and computer program capable of calculating an individual's credit rating.

According to an aspect of the present invention, there is provided a credit evaluation method comprising: setting a credit rating evaluation standard using a logistic regression model or a machine learning technique; , Analyzing the correlation between individual's psychological characteristics and credit risk to generate psychological evaluation criteria and measuring individual personality, attitude, mental health and basic capability And the psychological measurement data inputted as a response to the psychological measurement questionnaire is numerically expressed using an item analysis method or an average score calculation method, and based on the psychological evaluation criterion, A second evaluator for evaluating the psychological class using a logistic regression model or a machine learning technique based on the quantified psychological measurement data, Group comprises a first evaluation unit evaluated by the credit rating and to combine the psychological rating by the second evaluation unit for calculating an assessment final credit rating unit.

According to the present invention, since an individual's credit rating can be calculated based on a credit rating of an individual's information and a psychological rating of psychometric measurement data, individuals whose credit ratings are unfavorably applied due to lack of credit information can be easily It is possible to receive a loan audit and receive a loan without a security loan according to the calculated credit rating.

Further, according to the present invention, psychological measurement data can be created within a short period of time, so that the individual can minimize the psychological or temporal burden, and the financial institution can perform various kinds of research such as personal credit transaction, SNS related record, The result of the credit evaluation for the individual can be confirmed quickly regardless of the accumulation period or time.

Further, according to the present invention, since the credit evaluation is calculated by combining the characteristics of the individual, the willingness to repay the debt, and the repayment ability, the financial institution can obtain a credit evaluation result that is more refined and improved in discrimination about the individual. In addition, the financial institution can predict whether a default will occur based on the characteristics of the individual.

Further, according to the present invention, the financial institution can reduce the loan loss ratio while maintaining the loan approval rate or increase the loan approval rate while maintaining the loan loss ratio, based on the result of the individualized credit rating.

1 is a configuration diagram showing a personal credit evaluation apparatus according to an embodiment of the present invention;
Figure 2 is a flow chart illustrating a method for evaluating a credit rating in accordance with one embodiment of the present invention.
Figure 3 is a flow chart illustrating a method for evaluating psychometric data in accordance with an embodiment of the present invention.
Figure 4 - a flow chart illustrating a method of computing an individual's credit rating in accordance with one embodiment of the present invention;

The present invention relates to a psychological measurement data based personal credit evaluation apparatus, a method and a computer program. More specifically, the present invention relates to a personal credit evaluation apparatus, a method and a computer program, The present invention is applicable to all fields in which a credit evaluation for an individual is required. The personal credit evaluation device, method and computer program based on psychometric measurement data according to the present invention may be collectively referred to as Psychometrics-based Scoring System (PSS).

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a configuration diagram illustrating a personal credit evaluation apparatus according to an embodiment of the present invention.

1, the personal credit evaluation apparatus 10 according to the present invention may include an input unit 20, a credit evaluation calculation unit 30, and an output unit 40. [ The personal credit evaluation apparatus 10 according to the present invention may be a device operating in a financial institution to calculate a credit rating for an individual or may be a company or a public entity for providing a credit rating calculated for an individual as a financial institution Lt; / RTI > may be an apparatus included in a system operated by a user.

The input unit 20 receives predetermined information from an individual and receives information for evaluating a credit rating and psychological measurement data. The input unit 20 may be a keyboard or a touch pad that can be directly input by an individual. The input unit 20 may be a device that is connected to a computer, a tablet PC, a portable terminal, A financial institution or the like.

The output unit 40 is a device for requesting an individual to input certain information, and a list of psychological measurement questionnaires or a list of information to be collected for credit evaluation is output to the individual via the output unit 40. The output unit 40 may be a display unit that provides a screen directly to an individual, or may be a device that is connected to a computer, a tablet PC, a portable terminal, or the like used by an individual through an Internet network to transmit a list of information or a psychological measurement questionnaire, A list of information provided to an individual who visits the institution or the like, or a device in which a psychometric questionnaire is stored.

The credit evaluation calculating section 30 may include a first evaluating section 31, a second evaluating section 32 and an overall evaluating section 33. [ The credit rating calculator 30 can request the individual to input predetermined information through the output unit 40 and calculate the individual credit rating according to the information input through the input unit 20. [ The credit rating calculated by the credit rating calculator 30 may be provided to a financial institution. The first evaluation unit 31 evaluates the credit rating according to the input information, and the second evaluation unit 32 evaluates the psychological rating according to the input psycho measurement data.

In more detail, the first evaluation unit 31 sets a credit rating evaluation standard. The information related to the set credit rating evaluation standard is requested to the individual through the output unit 40 and the credit rating is evaluated according to the credit rating related information input from the individual through the input unit 30. [ The first evaluation unit 31 can receive basic information such as occupation and residential type from the individual, income information such as monthly disposable income, and can receive the credit information provided by a company such as NICE / KCB Can be received.

The first evaluation unit 31 uses a logistic regression method or a machine learning method to set a credit rating evaluation standard. Therefore, even when the number of evaluation items increases in the evaluation of the credit rating, the first evaluation unit 31 can maintain the predictive power at a high level and improve the accuracy and stability in the regression analysis. In evaluating the credit rating, the first evaluation unit 31 may calculate the credit rating as a digitized score, or may calculate the score by grouping the scores by intervals. Accordingly, the credit rating according to the present invention may be expressed as a score, or may be represented as a grouped grade by section.

The second evaluator 32 sets a psychometric evaluation criterion for the psychometric data and provides the psychometric questionnaire to the individual via the output 40. [ The second evaluating unit 32 evaluates the psychological class according to the psychometric measurement data inputted as a response to the psychometric questionnaire on the basis of the set evaluation criterion. In evaluating the psychological grade, the second evaluation unit 32 may calculate the psychological grade as a digitized score or may be calculated by grouping the scores by intervals. Therefore, the psychological class according to the present invention can be represented by a score, or can be represented as a step grouped by intervals.

First, the second evaluator 32 sets a psychological evaluation criterion by analyzing a correlation between an individual's psychological characteristic and an individual's credit risk. Two methods can be used to analyze the correlation between individual psychological characteristics and credit risk. First, a retrospective analysis method that confirms the correlation between psychological characteristics and credit risk by inputting psychological measurement data from a good loaner and a bad loaner among individuals who have received a conventional loan can be used. Second, A prospective method of analyzing the correlation between psychological characteristics and credit risk can be used by tracking the borrower's redemption records based on psychometric data obtained from new borrowers. In the former case, the accuracy of the correlation analysis decreases, but the correlation can be grasped quickly. In the latter case, it takes a long time to grasp the correlation, but the accuracy of the correlation analysis can be improved. The second evaluator 32 can check the correlation between psychological characteristics and credit risk using at least one of a retrospective analysis method and a prospective analysis method. If the two methods are used together, Correct correlation can be confirmed. The second evaluation unit 32 may use a logistic regression or a machine learning technique to analyze the correlation between psychological characteristics of an individual and credit risk.

A correlation between the psychological characteristic of the individual analyzed by the second evaluation unit 32 and the credit risk is generated as a psychological evaluation standard and the psychological grade of the individual is evaluated. The evaluation criteria can be verified through discriminant power and stability based on indicators such as K-S statistic, Area Under ROC Curve (AUROC), and Population Stability Index (PSI).

The second evaluator 32 may accumulate psychometric data for a plurality of individuals and statistically analyze the correlation between the psychological characteristics of the individual and the credit risk to provide a more reliable psychological rating have.

Next, the second evaluation unit 32 generates a psychological measurement questionnaire that can measure the psychological characteristics of the individual, and provides the questionnaire to the individual.

More specifically, the second evaluation unit 32 can classify the psychological characteristics of an individual into four categories: personality, attitude, mental health, and basic capability , And generates a psychometric questionnaire that can measure corresponding features.

In addition to the Big 5 Personality Theory, a subset of personality dimensions is used to measure the delay of gratification, impulsiveness, sensation-seeking, overconfidence, Self control, locus of control, self-efficacy, optimism, antisocial personality, moral judgment, and altruism can be analyzed .

Attitude to money, money management, attitude to debt, attitude to credit transaction, and materialism are the sub-items for measuring attitude. Can be analyzed.

Stress, depression, anxiety, and emotional well-being can be analyzed as sub items to measure mental health.

Sub-items to measure basic capability can be analyzed as management of money, coping strategy, intellectual capacity, and financial knowledge.

The personality, attitude, mental health, and basic competence that characterize an individual may vary according to psychology. Therefore, an analysis of intelligence can be additionally performed to supplement the analysis of psychological variables as described above. The intelligence can be analyzed using the Digit Span Recall test or the Ravens Progressive Matrix test.

Thus, the second evaluator 32 may generate a psychometric questionnaire that includes items for analyzing an individual's personality, attitude, mental health, basic competence or intelligence.

The second evaluation unit 32 may use a Likert scale, a forced choice item, a comparison evaluation item, and the like in the form of an item for evaluating the characteristics of the individual. In the case of the comparative items, the item is a combination of the items in the Riccart scale and the compulsory items. Thus, the psychometric survey questionnaire generated by the second evaluator 32 may include a Likert scale, a forced choice question, or a comparative question item.

Since the second evaluation unit 32 is not limited to the psychometric measurement based on the Riccart scale, the forced selection, or the comparative evaluation item, a picture or a game can be used for the psychological measurement questionnaire. Therefore, the psychometric questionnaire can be composed of a multiple choice, simple sentence or image as an option.

For example, the psychometric survey questionnaire can be used to analyze the psychological measurement questionnaire, such as personal value pattern, work life type, negotiation pattern, saving / spending pattern, rest pattern, consumption pattern, decision pattern, interpersonal pattern, The second evaluation unit 32 may include a question about the psychological change of the individual according to the change, the situation recognition pattern, the problem solving pattern, the emotion regulation pattern, etc., and the second evaluation unit 32 may provide the psychological measurement questionnaire to the individual .

One  Which of the following behaviors could be considered an adult? 2  How do you want to be recognized at work? 3  How do you start negotiating? 4  Is it a saving or consuming type? 5  Under what circumstances does the spending increase? 6  What do you do when you want to rest? 7  How often do you buy good stuff? 8  How much information do you need to change your mind? 9  What do you do after a big fight with a friendly person? 10  Which of the following images is the most annoying? 11  How did you feel when you heard that you were going to the moon? 12  What if I can not choose? 13  What does the first thing happen when something breaks in the house? 14  What type of party is it? 15  Do you know how much money you have in your account? 16  How does someone respond if you insult you? 17  What do you do when you have change? 18  What is the mood when things go wrong? 19  How serious do you play in the game?

The second evaluation unit 32 digitizes the inputted psychometric measurement data in response to the psychometric questionnaire. An item analysis method or an average score calculation method can be used to quantify an individual's personality score or grade. The item analysis method is a method of analyzing how well each item measures the characteristics of the individual to be measured and can provide high discriminative results. Item Response Theory, Classical Test Theory, and Rasch Analysis can be used as item analysis methods. Preferably Item Response Theory may be used. When using the item response theory to calculate the score, it is not a direct measure of the answer to the question, but it analyzes the potential factor that the answer is derived. Therefore, it minimizes the influence of the inspection environment and obtains high stability and reliability. . In addition, item response theory is a flexible model and has a higher practical value.

The second evaluating unit 32 analyzes the correlation between the psychological characteristics of the individual and the credit risk, and based on the psychological evaluation criterion generated by the psychological evaluation criterion, the second evaluator 32 uses the psychological measurement data quantified by the item analysis method or the average score calculation method, The psychological rating of

The comprehensive evaluation unit 33 combines the credit rating evaluated by the first evaluation unit 31 and the psychological rating evaluated by the second evaluation unit 32 to calculate the final credit rating. In the calculation process, a combination of credit and psychological grades may be used to construct a matrix or a stacking technique. The comprehensive evaluation unit 33 can calculate the credit rating more precisely by combining additional conditions with the final credit rating calculated by combining the credit rating and the psychological rating. For example, it is possible to combine conditions that can be overturned for a class corresponding to automatic approval of a loan or conditions for which a decision can be reversed for a class corresponding to an automatic denial of a loan. And the like may be used as additional conditions.

Therefore, according to the present invention, individuals having the same credit rating can be further subdivided, so that it is possible to more accurately predict whether an individual defaults on default.

2 is a flow diagram illustrating a method for evaluating a credit rating in accordance with an embodiment of the present invention.

As shown in FIG. 2, the first evaluation unit 31 receives information necessary for evaluating the credit rating through the input unit 20 (S50). For example, basic information such as occupation, type of housing, income information such as monthly disposable income, and credit information provided by a company such as NICE / KCB can be input.

The first evaluation unit 31 applies a credit rating evaluation standard to the input information (S51). Logistic regression or machine learning techniques can be used to establish credit rating criteria.

Next, the first evaluation unit 31 evaluates the individual's credit rating according to the input information (S52). The evaluated credit rating may not be disclosed to the individual. In addition, the first evaluation unit 31 may use various filtering to evaluate an individual's credit rating. For example, the first evaluation unit 31 may use the age, the provided credit rating, and the repayment rate as criteria for calculating the credit rating .

3 is a flow chart illustrating a method for evaluating psychometric data in accordance with an embodiment of the present invention.

As shown in FIG. 3, the second evaluation unit 32 generates a psychological evaluation criterion by analyzing a correlation between psychological characteristics of a person and a credit risk (S60). Logistic regression or machine learning techniques can be used to analyze correlations between individual psychological characteristics and credit risk. The psychometric evaluation criteria can be verified through discriminant power and stability based on indicators such as K-S statistic, Area Under ROC Curve (AUROC), and Population Stability Index (PSI).

The second evaluation unit 32 generates a psychological measurement questionnaire, provides it to the individual, and inputs the psychological measurement data through the input unit 20 (S61). Psychometric questionnaires can include items that measure an individual's personality, attitude, mental health, or basic competence, and may include items to measure an individual's intelligence.

For example, the psychometric questionnaire was designed to analyze the personality, attitudes, mental health, or basic competence of the individual, including the Big 5 Personality Theory, delay of gratification, impulsiveness, (eg, sensation-seeking, overconfidence, self control, locus of control, self-efficacy, optimism, antisocial personality, moral judgment, Attitude to debt, attitude to credit transaction, materialism, stress, depression, anxiety, depression, depression, depression, depression, depression, anxiety, emotional well-being, management of money, coping strategy, intellectual capacity, and financial knowledge. have.

In addition, the psychological measurement questionnaire may further include an item capable of analyzing an individual's intelligence based on a Digit Span Recall test or a Ravens Progressive Matrix test.

The psychometric measurement data entered as a response to the psychometric questionnaire may be collected based on the individual's consent.

Next, the second evaluation unit 32 converts the input psychological measurement data into a numerical value through an Item Analysis method or an average score calculation method (S62). Then, the second evaluation unit 32 compares the psychological rating of the individual according to the psychological evaluation criteria set in step S60 (S63).

4 is a flow diagram illustrating a method for calculating an individual's credit rating in accordance with one embodiment of the present invention.

The comprehensive evaluation unit 33 combines the credit rating evaluated in step S52 of FIG. 2 and the psychological rating evaluated in step S63 of FIG. 3 to calculate the final credit rating of the individual (S70). In calculating the final credit rating, a matrix technique or a stacking technique may be used.

The comprehensive evaluation unit 33 may calculate the final credit rating by combining additional conditions on the result of combining the credit rating and the psychological rating. For example, the final credit rating calculated by the comprehensive evaluation unit 33 may be classified into 1 to 10 grades. If the credit rating is calculated as 1 to 4 grades, , It can be set as the subject of precision examination and automatic rejection in the case of 8 to 10 grades. In the case of customers who are subject to close scrutiny, they can only be approved for loan if they pass the pre-determined examination criteria. If they meet the conditions for automatic approval high-side override , The credit rating may change, and if the auto-rejection target meets the Low-side Override criteria, the credit rating will be changed so that the rating corresponding to the automatic approval or the close examination And may be finally calculated.

The personal credit evaluation method according to the present invention is implemented as a processor-readable code on a processor-readable recording medium provided in a machine, a computer, or an integrated control device used by a financial institution, a credit evaluation company or a public institution It is possible. The processor-readable recording medium includes all kinds of recording apparatuses in which data that can be read by the processor is stored. Examples of the recording medium readable by the processor include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like, and also a carrier wave such as transmission over the Internet. In addition, the processor readable recording medium may be distributed over networked computer systems so that code readable by the processor in a distributed manner can be stored and executed.

As described above, the present invention relates to a personal credit evaluation apparatus, a method, and a computer program for calculating an individual's credit rating. By calculating a personal credit rating based on a new credit rating standard and psychological measurement data, Financial institutions can reduce the loan loss ratio to individuals or increase the loan approval rate because it is possible to classify the individual who wants to borrow more finely and predict more accurately whether an individual defaults.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the inventive concept as defined by the appended claims. But is not limited thereto.

Claims (8)

A first evaluator for setting a credit rating criterion using a logistic regression or machine learning technique and evaluating a credit rating for the information of the individual;
The purpose of this study was to analyze the correlation between psychological characteristics of individuals and credit risk to generate psychological evaluation criteria and to measure personality, attitude, mental health and basic capability And generating a psychological measurement questionnaire including a plurality of psychological measurement questionnaires and quantifying the psychological measurement data inputted as a response to the psychological questionnaire using an item analysis method or an average score calculation method, A second evaluator for evaluating the psychological class using a logistic regression model or a machine learning technique based on the quantified psychological measurement data; And
And a comprehensive evaluation unit for calculating a final credit rating by combining the credit rating evaluated by the first evaluation unit and the psychological rating evaluated by the second evaluation unit,
The second evaluating unit analyzes the correlation between the psychological characteristics of the individual and the credit risk. The second evaluating unit analyzes the correlation between the psychological characteristics and the credit risk by inputting the psychometric measurement data from the good lender and the bad lender And a proactive analysis method that confirms the correlation between psychological characteristics and credit risk by tracking the repayment records of borrowers based on psychological measurement data obtained from new borrowers Based on psychometric measurement data. The method according to claim 1,
Wherein the psychological measurement questionnaire includes at least one of a Likert scale, a forced choice item, and a comparison evaluation item in which a recall scale item and a compulsive selection item are combined. Based personal credit rating device. The method according to claim 1,
The psychometric measurement questionnaire further includes an item for analyzing the intelligence of the individual, and the items for analyzing the intelligence of the individual may include a Digit Span Recall test or a Ravens Progressive Matrix test, Based on the psychometric measurement data. The method according to claim 1,
Wherein the second evaluating unit verifies the psychological evaluation criterion based on a KS statistic, Area Under ROC Curve (AUROC), or Population Stability Index (PSI). The method according to claim 1,
Wherein the comprehensive evaluation unit constructs a matrix by a combination of the evaluated credit rating and the evaluated psychological rating or calculates a final credit rating using a stacking technique. A first step of setting a credit rating evaluation standard using a first evaluation unit logistic regression or a machine learning technique and evaluating a credit rating of the individual information;
The second evaluator generates a psychological evaluation criterion by analyzing the correlation between the psychological characteristic of the individual and the credit risk,
The items for measuring personality, attitude, mental health and basic capability of a person, and the Digit Span Recall test or Ravens progressive matrix test (Ravens) A psychological measurement questionnaire including an item for measuring an intelligence of an individual to be analyzed based on a progressive matrix test,
The psychological measurement data inputted as a response to the psychological measurement questionnaire is numerically expressed using an item analysis method or an average score calculation method and a score or grade is calculated through a logistic regression model or a machine learning technique,
The psychological evaluation criterion is verified based on KS statistic, Area Under ROC Curve (AUROC) or Population Stability Index (PSI)
A second step of evaluating the psychological grade using the quantified psychometric measurement data based on the psychometric evaluation criterion; And
And a third step of calculating a final credit rating by combining the credit rating evaluated by the first evaluation unit and the psychological rating evaluated by the second evaluation unit by a matrix or a stacking technique, ,
In analyzing the correlation between the psychological characteristics of the individual and the credit risk, the second evaluator receives the psychological measurement data from the superior loaner and the bad loaner among the individuals who have received the conventional loan, And psychological measurement data obtained from new borrowers. The results of this study are summarized as follows. First, we examine the relationship between psychological characteristics and credit risk by tracking the repayment records of borrowers based on psychological measurement data obtained from new borrowers. Personal data based on measurement data. The method according to claim 6,
Wherein the psychological measurement questionnaire includes at least one of a Likert scale, a forced choice item, and a comparison evaluation item in which a recall scale item and a compulsive selection item are combined. Based personal credit rating method. A computer-readable recording medium on which a computer program for performing the personal credit evaluation method according to claim 6 or 7 is recorded.

Images