Identification And Prevention Of Fraudulent Users In Internet Financial Platform

Due to the rapid development of technology,the concept of "Internet +" has become increasingly popular in recent years,which has changed the operation mode of many industries.The combination of the Internet and finance is particularly noticeable.We are pleased to see that because of the use of Internet technology to enable data sharing and information exchange,the "Internet + Finance" model breaks the barrier of information asymmetry in the tradit ional financial field,which is of great significance to China's economic development.However,internet finance is on the rise,but user fraud incidents are constantly emerging.As fraud methods become more professional,the risk of user fraud on Internet financial platforms continues to escalate.Therefore,how to identify fraudulent users in advance and make accurat e attacks to prevent them from becoming a problem has become an urgent problem in the current situation.This article starts with a study of the current status of fraud on Internet financial platforms,and analyzes the causes,development characteristics,and classification of fraudulent users.In order to explore the technical means of identifying user fraud,study the domestic and foreign literature on the detection of user fraud on Internet financial platforms,and sort out four types of user fraud recognition methods based on time series,relationship graph,biological probes,and data mining.In addition,this paper also uses the user transaction data of a European network payment platform as a sample to conduct empirical research on user fraud detection based on data mining.In this process,facing the problem of unbalanced positive and negative samples,a random under-sampling method is proposed to construct sub-samples,and then the sub-samples are processed for feature scaling,outlier detection,feature screening.The processed sub-samples are used to train the four classification models of Logistic Regression,K-Nears Neighbors,Decision Tree Classifier and Support Vector Classifier,and the prediction results of the four models are evaluated.The results show that the logistic regression model has the highest recall rate,F1 score and AUC value,indicating that the detection method based on data mining is feasible.In addition,based on the previous discussion on the causes of user fraud,this article proposes three directions for the prevention and control of fraudulent users of Internet financial platforms.They are to improve the level of Internet financial supervision,use fintech anti-fraud technology and build an order in the Internet financial industry.