Research On Effective Prediction Of The Internet Consuming Customers Of Financial Loan Products

With the rapid development of Internet finance,Internet financial products are born.Internet financial products mainly include online financial management,insurance and loan products.However,due to the imperfect domestic and personal credit information system,incomplete results of personal credit reporting,various opportunities and challenges of the personal credit industry,the situation is so complex that the customer identification of online credit products in China cannot select rapidly and accurately effective users with high willingness,low risk and good credit.Meanwhile,influenced by a series of factors such as macroeconomic development,the transform of consumers' concept,the convenience of online loan and the change of models of loan products,the speed of development of the credit loan business is so fast.Consumer credit loan has gradually become the current consumption habits of people and the mainstream method of people's living and consumption.Therefore,the development space of consumer credit loan in China is huge.According to the quick development of Internet consumer finance loan products,customer default and the adverse consequences of customer default occur from time to time.Only by accurately identifying and positioning the target customers,having a very precise understanding of the products,quantifying the credit situation of customers and the repayment risks of products,strengthening the risk management of personal consumption credit loans and reducing the risk of loss,the funds of company are safe.And the overall profitability can be guaranteed.Therefore,the study of this thesis has practical significance.In this thesis,we review the current situation and existing problems of credit market in China,Secondly,the development status and existing challenges of the personal credit industry are systematically summarized.Then,credit scoring cases about big data at home and abroad and two credit loan products from different companies are explored and analyzed.After that,comparing the credit scoring models from two countries and illustrating applications of effective customer identification of Internet financial loan products.Then,a purely online loan product of an internet financial company under a big A group is selected as this research object.Creating a model which is developed to effectively identify its target customers is a final purpose in this research.Before the establishment of the model,it mainly includes data preparation,data cleaning,data preprocessing,sample selection and so on.In the process of establishing the model,we mainly construct the logistic regression model and the data envelope analysis model.Then,using the data envelopment analysis model to pre-process the data,and select new indicators to be trained in the initial logistic regression model.Finally,the model accuracy,mixed matrix,high and low hit rate of user ratings and Kolmogorov-Smirnov test are used to test this model,and the effect of the model is also analyzed.After the development and research of this model,this thesis argues that the advanced big data technology should be used to collect,clean and process the massive multi-dimensional data of Internet financial loan product users,and then data envelopment analysis(DEA)is used to pre-process some indicators to increase the efficiency value of DEA in the traditional logistic regression model,so as to improve the accuracy of the model.The model results show that the short message response rate of the customers of the Internet loan product has been significantly improved,which makes the potential customers of the Internet loan product effectively identified.In short,the model can reduce the bad debt risk of Internet loan products and provide support for precision marketing and increasing revenue.