An Approach Utilizing Multiple Viewpoints to Identify Fraudulent Activity in Commercial Transactions Involving Multiple Parties

Identifying and preventing deceitful transactions on e-commerce platforms has persistently been the prime objective of transaction security mechanisms.However, owing to the covert nature of e-commerce, apprehending perpetrators solely based on historical order data is an arduous undertaking.Numerous studies have endeavored to develop technologies to thwart frauds, yet they have neglected to consider the dynamic conduct of users from multiple perspectives, leading to inefficient detection of fraudulent behaviors.To address this issue, this project proposes a novel fraud detection approach that amalgamates machine-learning and process mining models to monitor user behaviors in real-time.Firstly, we establish a process model concerning the B2C e-commerce platform, incorporating the detection of user behaviors.Secondly, a method for analyzing anomalies that can extract salient features from event logs is presented.Subsequently, we feed the extracted features to a Support Vector Machine (SVM) based classification model, which can detect fraudulent behaviors.Through experiments, we demonstrate the efficacy of our method in capturing dynamic fraudulent behaviors in e-commerce systems.