Design And Implementation Of Anti-fraud Risk Prediction System

In recent years,the development of the new financial model "Internet Finance" has attracted much attention.It breaks the barriers of the traditional finance and enables the general public to enjoy the services of inclusive finance.This model eases the imbalance between the supply and demand of funds,improves the efficiency of resource allocation,and increases the rate of financial penetration and utilization.However,there are many pain points behind the rapid development of the Internet Finance,such as credit risk,technical risk and regulatory risk.And the main content of credit risk is fraud risk.At present,the fraud industry is increasingly characterized by industrialization,concealment,specialization and sceneization.Therefore,for the financial lending industry,how to improve the anti fraud ability and the security of financial transactions has become an urgent problem to be solved.The purpose of this paper is to solve the fraud problem hidden in the credit process,so an anti-fraud risk prediction system is designed and implemented.The system can analyze the fraud risk of customer data,enhance the ability of financial institutions to identify and prevent fraud users,and effectively improve the risk control capability of the credit approval process.The system is mainly divided into two parts: system and model.The system is implemented in JAVA language,and uses Spring Boot,My Batis,Nginx and other technologies.It mainly realizes the functions of sample management,policy configuration,risk prediction and business monitoring.Model development is realized by Python language and uses machine learning technologies.The data set provided by the internship company and is balanced.And based on the data before and after the balance processing,we establish the logic regression model,xgboost model,xgboost + LR model,and finally choose the best xgboost + LR model.At present,the system has been successfully used,which satisfies the customer's need for fraud risk detection in credit business,and strengthens the ability to identify fraud risks for lending users,and ensures the borrower's safety qualification.