Research On The Influence Factors Of Personal Credit Consumer Loan Default Based On Big Data

In recent years,with the improvement of residents’ income level and the gradual formation of emerging consumer groups,consumer market in our country has developed spurt,and major Internet giants,P2 P platforms,consumer finance companies,and even traditional commercial banks have all laid out consumer financial markets.And quickly launched their own online and offline personal credit consumer loan products,personal credit consumer loans ushered in explosive development.However,at the same time of rapid development,horizontal competition and the loan demand for high-quality customer have gradually saturated,and the threshold for customers has been decreasing.As a result,the balance of non-performing loans and the ratio of non-performing loans have also risen steadily.The risk of default on loans has been continuously exposed.Therefore,it is urgent to identify and refine various factors that have a significant impact on the default of personal credit consumer loans and apply them to actual risk control,so as to promote a benign development track.This paper analyzes the current status of default risk control of personal credit consumer loans from the popular FICO personal credit scoring system,the default risk control of traditional commercial banks and the default risk control of emerging Internet financial enterprises.The credit information system is imperfect and the coverage of the population is relatively limited,which makes it impossible for financial institutions to effectively copy the popular FICO personal credit scoring model.While the traditional commercial banks still use the default risk assessment model to offset the pledge of personal credit consumer loans,the default risk assessment model cannot accurately reflect the potential default risk of customers.The credit evaluation system established by emerging Internet financial enterprises through big data and artificial intelligence technology also has defects such as data authenticity and validity.The actual application effect still needs to be tested by time.Based on this,from the perspective of theoretical analysis,this paper divides the influence factors of personal credit consumer loan default into two aspects: customer characteristic factor and loan characteristic factor,and extracts the customer loan application from the massive data of traditional commercial bank through big data technology.There are new characteristics such as the average daily financial assets held in the past six months and the number of related defaults,and there are few research results such as the number of historical loans,the frequency of monthly loans,and the proportion of credits actually occupied by customers’ loan applications.These characteristic factors are the focus of this paper.After comparing the Logistic regression model,the Lasso-logistic regression model and the decision tree model,the Logistic regression model with better prediction effect was selected to conduct an empirical analysis of the factors affecting the default of personal credit consumer loans.The accurate prediction ratio of the model is 81.34%,which indicates that the prediction model established in this paper is effective,and has a strong reference value and guiding value for the financial institution’s risk control of personal credit consumer loans.At last,the paper also proposes targeted suggest.