E-commerce Transaction Fraud Detection Using the Naive Bayes Algorithm

Abstract

This study utilizes the Naive Bayes algorithm to detect fraudulent transactions occurring on e-commerce platforms by analyzing several key attributes, including the transaction time, transaction amount, the user's geographic location, and the payment method used. This algorithm was chosen due to its advantage of simplicity in handling probabilistic-based classification, which facilitates the analysis of complex data. Based on the study's findings, the Naive Bayes model demonstrates a commendable ability with an accuracy rate of 80% in identifying transactions categorized as fraudulent activities. This research contributes valuable insights that can be applied to enhance the security and trust in online transaction systems.