Customer Credit Risk Rating Reduction and Scientific Financial Planning
Download as PDF
DOI: 10.23977/EMCG2020.008
Author(s)
Liu Ziyang, Huo Zhaodong, Wang Linmiao
Corresponding Author
Liu Ziyang
ABSTRACT
In this paper, we first analyze the customer's capital status based on the Bank flow data, establish a standard to classify the credit risk, and screen out the customers with high possibility of money laundering risk. Then, combined with the annual interest rate of four products of loan, deposit, installment and financial management, we provide reasonable choice suggestions for high-quality customers, and predict the change trend of the four products in the next three years. Finally, the sensitivity and stability of the model are tested. Aiming at the problem of customer credit risk classification, we use risk gradient reduction model for qualitative analysis. Firstly, two concepts of default principal and default probability are defined, in which default principal represents customer value to a certain extent. TOPSIS Model Based on entropy weight method is used to calculate the scores of various users. In view of the problem of product portfolio recommendation, starting from the net transaction volume n between customers and banks in 2018 and considering the capital status of customers, in order to make the proposal more tendentious, we propose a propensity utility function, which divides the portfolio into two types: inward investment and outward credit, and then combines the two indicators of transaction frequency and relative conversion rate, as well as the four products of loan, deposit, installment and financial management. The interest rate situation provides targeted portfolio suggestions for customers respectively.
KEYWORDS
Utility maximization, RFM two dimensional standard, risk gradient reduction