Machine Learning For Website Defacement Detection: A Survey Of Techniques, Trends, And Challenges

Abstract

Web defacement attacks are rapidly changing cyber attacks, characterized by unauthorized alteration of online content and misleading techniques utilized to trick users. The rate of cyberattacks is on the rise globally reflected by nearly 600 cases reported in India in the first half of 2024 implying that conventional defense tools are slowly losing their effectiveness. Against this backdrop, Machine Learning (ML) has emerged as a powerful tool for identifying and combating these intrusions. This survey provides a comprehensive overview of recent advances until 2025, with ML-based approaches to web defacement detection and associated threats like phishing URLs and malicious activity. Following seminal and recent work, we compare a variety of ML methods, from traditional algorithms to deep learning and ensemble methods, on the basis of accuracy, scalability, and usability. Further, the paper emphasizes existing challenges, including data imbalance, adaptive attack techniques employed by attackers, and the necessity of real-time detection, while emphasizing emerging trends and possible avenues for future research.