Online lender credit-rating method based on factor score K-Means clustering

Abstract

The invention discloses a online lender credit-rating method based on factor score K-Means clustering and the method comprises the following steps: S1, obtaining sample data, preprocessing and standardizing the sample data; S2, outputting a KMO statistical magnitude and a sphericity test significance value of Bartlett by applying the standardized sample data in the step S1, and determining sample data suitable for factor analysis; S3, outputting common factor variances of the sample data suitable for factor analysis, extracting common factors in the common factor variances by using the factor analysis, explaining the meaning of each common factor by using a rotation component matrix, naming the common factors, and finally calculating factor scores according to a factor score matrix; S4, performing clustering analysis on the sample data through a K-Means clustering algorithm; and S5, introducing a default recovery rate index as a quantitative standard for credit rating judgment. According to the invention, credit risks of borrowers at different levels can be reasonably distinguished, and the credit rating accuracy is high.

Description

technical field [0001] The invention relates to the field of information technology, in particular to a credit rating method for online loan borrowers based on factor score K-Means clustering. Background technique [0002] Since 2013, the online lending industry has developed rapidly in China. However, as the scale of the industry continues to expand, various problems have gradually been exposed. A high proportion of problematic platforms and thunderstorms have occurred frequently. P2P online lending institutions, which mainly focus on more problems, will also completely return to zero in mid-November 2020. At present, the entities engaged in online lending include 1) microfinance companies, which usually do not have financial licenses; 2) consumer finance companies, which are mainly bank subsidiaries separated from banks, and hold financial licenses like banks; 3) commercial banks, relatively Strict supervision and formal operation. [0003] However, while strictly managi...

AI Technical Summary

Problems solved by technology

In contrast, in China, even online lending platforms such as Renrendai, which have been officially connected to the basic database of financial credit information of the People's Bank of China (that is, the credit information system), still have major problems in the credit rating process of online loan borrowers. , the credit scoring standard adopted by Renrendai, as long as the borrower fills in the information as required, the credit score can be assessed, and the more additional information is provided, the higher the credit score is, and the past repayment records will also affect the credit score. Substantial impact

After the credit score is formed, Renrendai is divided into seven credit ratings of AA, A, B, C, D, E, and HR according to the score range it belongs to. Each credit rating has a different service fee rate, but the data statistics It shows that the credit rating mechanism of Renrendai does not play a role in differentiating borrowers from the perspective of default recovery rate

[0005] The existing disclosed patents related to online loan credit rating or the credit rating method of online loan borrowers that have been implemented have the following deficiencies: mainly based on the overall online loan platform as the credit rating object, the rating established based on the existing domestic credit data II It is impossible to reasonably distinguish the credit risks of borrowers of different grades, and instead, there are a large number of borrowers with high credit ratings and a high default rate

Images