Credit Risk Prediction Method Of Chinese Listed Companies

Credit risk prediction is using the financial data and non-financial data of the company to predict the corporate financial situation in the future.Credit risk prediction is important for investors and creditors to make decisions because it can help them identify credit risk.No matter the trading of stocks,the investment of corporate bonds,the issuance of bank loans or the business credit granting between upstream and downstream enterprises,credit risk prediction and estimation are of vital importance.This paper studies the credit risk prediction method of Chinese listed companies.The first chapter is the introduction,the second chapter is the theoretical basis of the credit risk prediction model,the third chapter is the construction of the credit risk prediction model,the fourth chapter is the empirical analysis of the credit risk prediction model of Chinese listed companies,and the last chapter is the conclusion.The focus of this paper:The first is the determination of the standard of accuracy to build the prediction model.If the standard of accuracy does not pay attention to control the Type II error,it will lead bad companies to be misclassified as good companies,which may cause credit risks and crises.The second is the construction of credit risk prediction model.Obviously,the prediction accuracy of different models is different.The third is the determination of the model’s prediction preiod.The prediction period is a key measure of the predictive power of the model.There are two main innovations in this paper.The first is to obtain weighted misclassification rates by using different values of C_k,which is the ratio of coefficient of Type II error and Type I error,and different values of t_q,which is the threshold of linear discriminant equantion.Then deducing the optimal weight W_k~*and the optimal threshold t_k~*corresponding to different C_k by minimizing weighted misclassification rate and deducing the optimal weight W~*and the optimal threshold t~*by maximizing AUC of the prediction models under W_k~*and t_k~*.The linear discriminant equation constructed with this method is the credit risk prediction model of this paper.Second,the prediction accuracy of model at the time T+d in this paper is higher than that of typical linear discriminant analysis,logistic regression,support vector machine,naive bayes and K nearest neighbor models,and the prediction period d is larger than that of existing studies.The study found that the credit characteristics of the geographical regions of Chineses listed companies are the best for the listed companies in East China and South China;the credit qualifications in North China,Central China and Southwest China area round the average,while the credit qualifications in Northeast China and Northwest China are the worst.The industry credit characteristics of Chinese listed companies are as follows:“Financial Industry”,“Information Transmission,Software and Information Technology Service Industry”and“Scientific Research and Technology Service Industry”have the best credit qualifications,while the“Agriculture,Forestry,Animal Husbandry,Fishery Industry”,“Real Estate”and“Mining Industry”are the worst.