The Application Of RF In Credit Risk Early Warning Of Listed Companies

With the rapid development of Chinese securities market,more and more companies are trying to be listed to obtain capital,so that they can enhance their financial strength in order to expand their business scope and occupy a position of prominence in market competition.For Chinese economic development,Listed companies have been the core driving force.However,due to various reasons,some listed companies may fall into credit crisis and face the situation of operating losses,and some listed company even been marked ST or *ST by regulatory authority.The emergence of credit defaults of listed companies will bring huge losses to stakeholders.Therefore,It has important and practical significance to construct an effective credit risk warning model for listed companies.This paper takes the industrial industry in the wind industry classification standard as the research object.this paper make studies from two aspects,which are the construction of credit risk evaluation index system and the establishment of credit risk assessment model.In the construction of random forest model,this paper optimizes the number of decision trees and the number of indicator variables in internal node to improve the prediction accuracy of the model.This paper selects 23 listed companies in the industrial industry that have been marked ST or *ST in 2018,and selected 69 normal listed companies in the same industry according to the ratio,and use annual financial data for the three years from 2015 to 2017.The samples are divided into training set and test set according to the ratio of 3:1,the former is used to train the model,and the latter is used to verify the warning effect of the model.Finally,the prediction effect of the random forest is compared with the prediction effect of the CART decision tree,It shows that the prediction effect of random forest is better than CART decision tree.