The Fraud Supplier Identify Mechanism In E-commerce Based On BP Neural Network

With the rapid development of electronic commerce, the asymmetry of informationbetween the sellers and buyers becomes more obvious in the network market than in thetraditional one. The fraud problem caused by the the asymmetry of the product information,not only exists but also becomes more serious. E-commerce as the dominant of the virtualeconomy is the important component of the modern economy. However, since laws andregulations in this aspect in our country are still not perfect, the virtual economy still bringsrisks and hazards to some degree. Therefore, the realization of the monitoring and commercialevaluation of the E-commerce transaction risk, actively prediction, risk-analysis, reducing theuncertainty of the transaction, has become an urgent need of the development of electroniccommerce. This paper attempts to analyze the fraud in E-commerce suppliers, combined withasymmetric information, BP neural network, business intelligence and other related theory ofsupplier fraud warning.First of all, in the basis of fraud pattern analysis on the traditional industries, asymmetricinformation theory is used to analyze the classification, characteristics and the development ofE-commerce fraud, and the game theory is also used to analyze the evolution. Secondly,through the research of trust model, according to the characteristics of e-commerce system, asuppliers trust model is constructed.The user behavior and relationships of the supplier are analysed to mine the hiddenbehavior and the relationship with the fraud users, such as IP Association, E-maill associationand so on. According to the above fraud behavior inference, complete e-commerce supplierfraud warning mechanism is built. From three aspects of static information, dynamic behaviorand its correlation, all-round, multi-dimensional analysed the fraud, so as to further improvethe accuracy and coverage of fraud warning. Again, BP neural network recognition algorithmis introduced to implementation of self-learning, adaptive and supplier fraud warning model.Finally, through the analysis of the characteristics of the the case company’s data and theexisting electronic commerce systems, the supplier fraud warning business intelligence systemis designed. Based on the data integration, the reasonable BI system architecture, datawarehouse, ETL, multi-dimensional analysis and scheduling strategy are designed to make itapplicable to electronic commerce system.In this paper, the behavior analysis, correlation analysis is introduced into the design ofthe warning model of E-commerce supplier fraud. Also the BP neural network identification isused to improve the learning and adaptation. With the expansion of the data scale, complexchanges in fraud factors, the effect would be more obvious, especially when the scheme is introduced into the electronic commerce environment. Through the simulation test of theaccuracy and consistency, the feasibility and superiority of the model are proved comparingwith the other ones. This study laid a foundation of theory and technology for the commercialrealization of the supplier fraud warning in E-commerce.