Construction of Bank Credit White List Access System Based on Grey Clustering Algorithm

Abstract—While consumer credit report is used as important reference for lending institutions to make lending decisions, personal credit score is regarded as the key tool to make decisions quickly and accurately. Problems arise in the utilization of traditional clustering algorithm to deal with credit risk problem, as a large number of redundant features may cause clusters to be distributed in a certain feature subspace of high-dimensional space, and the sparse high dimensional space makes the similarity of samples difficult to measure. In order to address the issues, this paper constructs a bank credit white list access system based on the grey clustering algorithm. This paper first lists all the factors that may affect personal credit, and makes a qualitative analysis one by one, and then screens these factors. When the characteristic variables of a new applicant are known, this discriminant function is used to predict whether the applicant is a good customer, and then decide whether to grant a loan to it. Combine multiple source data into a consistent data store by using data integration technology, and then store it as a data cube; Finally, data reduction and aggregation are used, redundant features are deleted, and data is compressed to obtain effective and usable data. In this paper, the accuracy of constructing bank credit white list based on grey clustering algorithm is 92.14%. The research results show that the model is effective.