KR19990019276A - Household Customer Evaluation System - Google Patents

Abstract

The present invention comprehensively analyzes accumulated data of bank household customers (for example, variables such as age, income level, housing situation, and transaction performance) to measure and accept risks of variables and systemize them to provide fast loans to customers. The purpose is to provide a decision and reasonable credit limit.

In order to realize this, the present invention uses the independence test to measure the risks for the variables, and then calculate scores for each variable using the maximum likelihood estimation method as a parameter estimation method in logit analysis, and create a scorecard, In addition to the decision criteria and credit limit setting criteria by income, the computerized system enables scientific and automatic screening of household customers.

Maximize customer satisfaction by quickly processing small-scale large-scale household loans to unspecified people through the use of the system, and improve the efficiency of business, and banks can select good and bad customers, thereby reducing household loans. Minimize effect can be obtained.

Description

Household Customer Evaluation System

The present invention comprehensively analyzes accumulated data of bank household customers (for example, variables such as age, income level, housing situation, transaction performance, etc.) to measure and score risks for variables to quickly determine loans and rationality to customers. It is a computer system that provides a credit limit.

The prior art in this field has failed to comprehensively analyze the data of bank household customers (such as variables such as age, income level, housing conditions, and transaction performance), so that screening based on the risk-tested criteria or the experience of the loan officer is not possible. In this regard, loan determination and loan execution are decided according to the professional ability and experience of the person in charge, and it takes a lot of time because of the characteristics of household loans in which a small amount of large loans are made.

After measuring the risks of these variables using the independence test technique, the score analysis for each variable was used as a parameter estimation method in logit analysis to calculate scores for each variable. It is a computerized system with credit limit setting standards to enable scientific and automatic screening of household customers through computerization.

1 is a flow chart of household customer evaluation system.

2 is a flow chart for determining loan.

1 is a flow chart of a household customer evaluation system, which is based on historical data (2) accumulated earlier, data on age, housing status, marital status, occupation, position, working years, annual income, number of accompanying family members, credit card possession, etc. Analyze (3). The technique used at this time is an independence test, an analysis method that checks correlation between variables. For example, if passive risks are not related to each other, the two variables can be argued for independence, and income cannot be used as a variable to describe risk.

If income and risk are related to each other, one can conclude that two variables are dependent and income can be used as a variable to explain risk.

The following is the scorecard item selection and scoring (4). The items with high risk are selected from the items that can be used as variables, and the weight is calculated by weighting each variable using the maximum likelihood estimation method as the parameter estimation method in logit analysis.

In this way, the scorecard preparation (5) is made and the monthly income for each score is matched to establish a reasonable credit limit (6). In other words, even if the score is the same as the superior customer with higher score, the limit is more limited to the customer with higher monthly income, and the credit is limited to the customers and the creditors who are below the standard score.

The scorecard cannot be continuously used as the bank's external environment and customer types change. When new data (1) is accumulated, the scorecard is updated every three to five years through risk reanalysis.

When the branch user (7) receives a loan application form and confirmation document from the customer upon request of the customer's loan, and inputs all the details of the loan application into the terminal, the computerized host (8) requests the age, occupation, and income among the customer's information items. The system operation (9) automatically scores the items such as the number of family members, and the items such as the banking record, the receiving balance, payroll, direct debit, and the presence of a credit card. The scorecard output (10) is performed. The provisional determination 11 is performed and this is demonstrated concretely in FIG.

Fig. 2 is a loan deciding flow chart, in which the counter clerk inputs the details of the loan application collected from the customer to the terminal 1 and outputs the scorecard 2 and the bad trader information linked to the bank association. The loan is decided by computerization.

In the computerized decision (4), the scorer who scores below the reference point and the current credit information defector etc. are automatically rejected (5), and those who are higher than the standard score are automatically approved (6) and the notification of rejection (11) or approval notification (12).

In addition, the matters in which it is difficult to decide whether or not to make a decision by computer are subject to review (7) but it is less than the standard point. 11) or the approval notice 12, and if there is no consideration, the direct notification (11) without going through the final review (10) process.

If there is any consideration of loan application refusal (9) such as bad credit information in the past, it is passed through the final review (10) of the creditor's predecessor (10). (10) Direct approval notification (12) without going through the process.

Claims (2)

A technique for creating scorecards by analyzing accumulated customer historical data Scientific and automatic screening system by linking credit card matched with score card, credit decision system and monthly income by score