Method used for constructing and verifying credit scoring equation and system design

Abstract

The invention discloses a method used for constructing and verifying a credit scoring equation and system design. A central computer server is connected with a public network. The central computer server is provided with a computer usable medium based on a series of instructions executed by a processor. The electronic process for the processor to assess the credit risk of a borrower comprises the steps that 1) the data set of the borrower are searched and collected from at least one data source through a public network; 2) the data set is converted into variables associated with the credit risk of the borrower; 3) each variable is independently processed to generate a meta-variable describing the specific aspect of the borrower by statistical or machine learning; and 4) the target credit risk score is calculated based on a number of variables and meta-variables of the borrower. According to the invention, each variable in various variables is independently processed to produce a series of independent decision sets which describe the specific aspect (meta-variable) of the borrower.

Description

technical field [0001] The invention relates to the technical field of credit scoring, in particular to a method and system design for constructing and verifying credit scoring equations. Background technique [0002] It is widely used in people's daily purchase behavior. In the United States in the 1950s, credit decisions were made by bank credit reviewers. Usually, since the credit reviewer and the applicant lived in the same area and were familiar with the applicant, the credit reviewer decided whether to grant a loan based on what he knew about the applicant. Although effective, this method is very limited because the number of credit examiners is always much smaller than the number of applicants. In the 1970s, the emergence of FICO scores greatly facilitated credit approval, effectively reducing the reliance on credit reviewers in the credit approval process. However, the function of risk control is still incomplete. Lenders such as banks and credit card companies us...

AI Technical Summary

Problems solved by technology

There is no traditional variable conversion method that can simultaneously consider hundreds of fields (thousands, tens of thousands, or even less at the million level)

Images