The Credit Risk Of Listed Companies Based On SVM Identifying Research

Listed companies are an important part of the securities market,the financial situation is good or bad with the development of the securities market and the interests of the majority of investors are closely related.In recent years,with the development of economy,listed companies face many opportunities and also many risks at the same time,the financial credit risk is the most prominent,become a hot research in recent years.Credit risk research methods development from the traditional qualitative analysis,statistical analysis methods to the machine learning and artificial intelligence and other methods.Practice has proved that support vector machine in credit risk research highlights many advantages.This paper is based on this idea,starting from China's listed companies,to study the enterprise's credit risk.In the past,the method of using support vector machine as a tool to study credit risk,at the multi-dimensional data processing generally use of principal component analysis,factor analysis and other linear dimensionality reduction method to reduce the dimension,but also the introduction of rough set method.In this paper,the dimensionality reduction method can be improved,and the nonlinear dimensionality reduction method can be introduced.In this paper,16 indicators of corporate solvency,operation ability,profitability and development ability are selected from the financial indicators of listed companies.Then,a principal component analysis method,multidimensional scale method,local linear embedding method and isometric mapping method are used to reduce the dimension of a large number of variables,and a compact input matrix is obtained.Finally,this paper uses the support vector machine(SVM)method to establish the credit judgment model of listed companies,and compares the advantages and disadvantages of the support vector machine model under different dimension reduction methods.