Research And Implementation Of Credit Card Overdue Risk Prediction Based On Deep Learning

With the explosive growth of credit card issuance in recent years,the current credit card business has brought profits to the commercial banks,but the high profits also brought a high risk.Credit risk has always been an important risk facing the banking industry,which has a great relationship with the development of the commercial banks.Therefore,how to effectively identify credit card customers,reduce the credit limit of high risk credit card customers,in order to reduce the loss of credit risk to commercial banks,is the focus of the prevention and control of credit risk.Based on the difficulties of the current credit card overdue risk prediction and the excellent feature learning ability of Deep Learning,the paper puts forward a method of predicting the overdue risk of the credit card based on the Deep Learning.In this paper,based on the full analysis of the data warehouse of a commercial bank,this paper firstly analyzes the derived data.After Analysis,deletion and standardization of input items,achieve a Deep Learning based classifier model for predicting the risk of overdue credit card of individual customer.After tuning the model,the fine-tuning of the depth of the neural network structure,taking into account the performance of the machine in the experimental environment case as far as possible to optimize the experimental results.Aiming at the problem of over fitting of the model,this paper uses the more popular over fitting solution in recent years to deal with,and does a test contrast of the effect of these solutions.