Online Fraud Detection Model Based on Social Network Analysis

摘要

With the rapid development of the Internet, the way of our living and thinking has changed. Because of the anonymity and low-cost legal sanctions, e-commerce has been booming on the Internet. Unfortunately, rapid commercial success has made e-commerce sites a lucrative medium for committing fraud. Therefore, we proposed a method that used for fraud detection and prevention on platform. First, we implemented a parallel web crawling agent to collect real users and transaction data. Second, we proposed Reverse Graph and Common Trade Cumulative Graph (CTCG) theory to extract features of common transaction. Third, we extracted the features of graph-level based on the Page-Rank and K-core clustering algorithm and replaced the PageRank values with reasonable TrustRank values, added BadRank values for identifying potential fraud users. Finally, we conducted a series of experiments using the Random Forest and verified the performance of our method by applying it to real transaction cases. In summary, our proposed model is effective in identifying potential fraudulent users on the fraud platform.