Machine Learning for Formjacking Attacks on Checkout Pages through Real-Time Transaction Analysis

A major risk for e-commerce sites and their victims alike, formjacking is a cyberattack technique that injects malicious code into web pages and captures sensitive user input. A Machine Learning (ML) based approach is proposed in the paper that analyzes transactions in real time to alert users if formjacking is present on checkout pages. The ML methods focus on anomaly detection, supervised learning for transaction frequency patterns recognition, and unsupervised clustering, and shows how it can effectively detect suspicious transactions and identify risks for formjacking attempts. An adaptable and scalable web model is also presented, which can be incorporated into the pre-existing web infrastructures in order to strengthen the security of online transactions.