Research On Enterprise Bankruptcy Prediction Based On Data Mining Technology

The most important job in bank credit activities is to control credit risk.In the process of financing loans to enterprises,it is necessary to analyze the operating conditions of the companies,especially to assess whether there are bankruptcy risks.Using artificial intelligence,big data analysis and data mining techniques,a variety of different enterprise bankruptcy prediction models can be established.The main tasks of this article are:(1)A bankruptcy prediction model based on the support vector machine method is proposed,and the model is compared with the enterprise bankruptcy prediction model based on the back propagation neural network method.The experimental results show that the enterprise bankruptcy prediction model based on the support vector machine method has higher prediction accuracy and better generalization performance.(2)An enterprise bankruptcy prediction model based on extreme learning machines and financial knowledge is proposed,which is an incremental extreme learning machine method for retention rate.The experimental results show that the 9 variables selected through this model are better than the 12 variables selected through simple financial knowledge.(3)A bankruptcy prediction model based on the kernel extreme learning machine method is proposed.Experimental results show that the performance of this model is superior to the extreme learning machine model,the support vector machine model,and the random forest model in terms of type I error,type II error,accuracy,and area under the characteristic curve.Therefore,this model is the most promising alternative to the enterprise bankruptcy early warning system.