Research On The Construction And Method Application Of Comprehensive Evaluation Index System Based On Broker Rating

With the rapid development of China's economy and the continuous establishment of economic systems,in order to ensure steady economic de-velopment,prevent credit risks,and maintain a normal economic order,the importance of credit risk assessment has become increasingly evident,and studies on credit risk assessment have been intensified.It has important practical significance.This paper first summarizes the research status of credit risk assess-ment at home and abroad,and elaborates the basic theory of credit risk and commonly used credit risk assessment methods.Again,the theoretical ba-sis of the empirical analysis of this paper,the C4.5 algorithm and the SVM algorithm The principle is summarized,and then a set of credit risk assess-ment index system is constructed based on the construction principle of the credit risk assessment index system.The credit risk assessment index system covers a total of 24 indicators in 5 areas and basically covers all factors that affect the credit risk assessment of listed companies.Finally,using the mixed modeling method of C4.5 decision tree algorithm and SVM support vector machine algorithm and the 6 fold cross validation method,the scikit-learn machine learning package under Python platform was used to train and ad-just the model,and finally obtained.The results of the model with the best prediction effect were used.Finally,the results of the C4.5-SVM model were compared with the traditional multivariate discriminant model and Logisitc model results using ROC curves and AUC values.The empirical analysis results show that the hybrid model combining the C4.5 decision tree algorithm and the SVM support vector machine al-gorithm has a higher prediction effect than the multiple discriminant model and Logisitc model.Therefore,the C4.5-SVM model proposed in this paper can effectively assess the credit risk of listed companies in the manufacturing industry.