Customer Credit Evaluation Based On Bayesian Classification Method

Credit evaluation classification is an important part of enterprise risk management.Its main objective is to identify good customers and bad customers according to customer information,find out the key influencing factors of customer credit,and provide references and suggestions for the credit decision-making of enterprise management.In order to correctly distinguish different customers and give consideration to the practicability of the enterprise in the process of credit evaluation,this paper adopts the Bayesian classification method to classify the credit status of the enterprise customer in terms of the credit data.In this model,the sample data is first preprocessed,the credit evaluation indicator system including non-financial factor variables is screened out,then the Bayesian network structure method is selected,the learning parameters are determined,and the preprocessed data is classified and analyzed.For verification,under the same condition,it is compared with other credit evaluation classification methods.The results showed that the accuracy,sensitivity and specificity of the tree-augmented Naive Bayes(TAN)model were the most accurate under the condition of 10-fold cross validation.At the same time,through the analysis of the results,it can be concluded that the nature of capital,registered capital and business reputation are the key factors of influencing the credit of enterprises.This means that these key variables are important indicators that need to be reviewed in the process of business operation and risk control.