A MULTI-PERSPECTIVE FRAUD DETECTION METHOD FOR MULTI-PARTICIPANT E-COMMERCE TRANSACTIONS

Abstract

Detection and prevention of fraudulent transactions in e-commerce platforms have always been the focus of 
transaction security systems. However, due to the concealment of e-commerce, it is not easy to capture attackers solely 
based on the historic order information. Many researches try to develop technologies to prevent the frauds, which have 
not considered the dynamic behaviors of users from multiple perspectives. This leads to an inefficient detection of 
fraudulent behaviors. To this end, this paper proposes a novel fraud detection method that integrates machine-learning 
and process mining models to monitor real-time user behaviors. First, we establish a process model concerning the 
B2C e-commerce platform, by incorporating the detection of user behaviors. Second, a method for analyzing 
abnormalities that can extract important features from event logs is presented. Then, we feed the extracted features to 
a Support Vector Machine (SVM) based classification model that can detect fraud behaviors. We demonstrate the 
effectiveness of our method in capturing dynamic fraudulent behaviors in e-commerce systems through the 
experiments.