Fraud Detection in E-Commerce Transactions Using Machine Learning and AI

With the rapid growth of e-commerce platforms, online transaction fraud has become a major concern for businesses and consumers alike. Traditional rule-based systems are increasingly inadequate in detecting evolving and sophisticated fraudulent behaviors. This study explores the application of machine learning techniques to detect fraudulent e-commerce transactions with higher accuracy and efficiency. We analyze a range of supervised learning models, including Logistic Regression, Decision Trees, Random Forest, and Support Vector Machines, comparing their performance using key metrics such as precision, recall, F1- score, and accuracy. A publicly available dataset simulating e- commerce transactions is used to train and test the models. Our results show that machine learning methods can significantly improve fraud detection rates while reducing false positives. The paper also highlights the importance of data preprocessing, feature selection, and handling class imbalance to optimize model performance. This research demonstrates the potential of machine learning as a reliable and scalable approach to enhance the security of online transactions in the e-commerce sector.