Commercial Bank Credit Risk Assessment And Dispersion Technology Research

China's banking will be confronted with more severe credit risks after enteringto the WTO. Firstly, domestic and overseas commercial banks will strive keenly forgood clients. With decreasing of good clients, domestic commercial banks had to givecredit to clients whose returns or credit grades are low, which exacerbate the creditrisks. In addition, except for financial industry, other organizations are in face of astrong competition with foreign corporations, which make enterprises risks and creditrisks worsen. Therefore, the issue of measure and diversification of credit risks incommercial banks is drawing much attention and is being extensively studied.This dissertation is mainly made from two aspects. On the one hand, apply fuzzymathematics to deal with problems involved with credit risks.1. Based on fuzzy set theory, establish different membership functions (MFs) forspecial financial targets to make the targets data standard by calculation andcomparison. After introduce the concepts of fuzzy entropy, derive entropy weightsfrom it to modify experimental weights of targets, and gain combined weights,consequently set up target system.2. According to the professional practices most of the banks in the internationalbanking adopt in, the credit risk assessment includes two aspects: an obligor's creditrisk assessment and the corresponding loans program credit risk assessment. Basedon fuzzy set theory, develop a distance discriminant model to evaluate the obligor'scredit risk. Afterwards, establish 0-1 integer programming model to evaluate the loansprogram credit risk, thereby optimizing the loans portfolio and solving the issue onoptimal decision-making of loans portfolio.On the other hand, introduce CCA/DEA to evaluate obligors'credit risk. Dateenvelopment analysis (DEA) is a new technique for measuring relative efficiencyamong the decision making units (DMUs) with multiple input and multiple output. Itis becoming one of the fast expanded methods in management science andoperational research over the last decades. Canonical correlation analysis (CCA) is anextension of regression analysis. Canonical correlation analyzes the relations of twogroups of variables. As a new technique, the application of DEA has limitation.