Risk Quantification of Small and Medium-Sized Enterprises and Bank Optimal Credit Strategy Model
Download as PDF
DOI: 10.23977/GEBM2020.001
Author(s)
Hu Hanqing, Du Jian, Bi Gaoang
Corresponding Author
Hu Hanqing
ABSTRACT
Based on principal component analysis and BP neural network model, this paper constructs a quantitative model of enterprise loan risk, and solves the optimal loan strategy of different types of enterprises according to the idea of game theory and nonlinear programming model. Given the credit rating and default data of an enterprise, the number of upstream enterprises, the total input amount, the total input tax, the number of downstream enterprises, the total output amount, the total output tax, and the output void rate are obtained through data cleaning and mining. Combined with the data of rating and default, we get the average score of each enterprise by principal component analysis. In the case of no given enterprise credit rating and default data, according to the possible internal relationship between enterprise indicators and credit rating, a BP neural network model is established. Considering the impact of emergencies on enterprise risk and bank credit strategy, this paper uses the method of text analysis to classify enterprises into corresponding industries. The framework of each industry risk model is established by using Delphi method, and the industry risk value of corresponding enterprises is modified, and the optimal loan amount and the optimal interest rate under the influence of industry risk are obtained.
KEYWORDS
Bank credit decision model, risk quantification, BP neural network, nonlinear programming, game theory