A Recent Study of Machine Learning Based Techniques for the Detection of Cyber-Attacks on Web Applications

Detecting cyberattacks remains a challenging task in almost every digital environment, particularly when dealing with passive attacks. Major organizations have suffered from financial losses and negative publicity due to the lack of timely detection of ongoing intrusions. According to IBM Analytics 2022, 83% of companies are at risk of falling victim to cyberattacks. This can happen at any time, and the shocking fact is that in 2022, just in the United States, the typical expense of a data breach was nearly 9.5 million US dollars. This paper provides a comparative analysis of researchers' methodologies and architectures proposed in recent years to detect and mitigate this issue. Furthermore, the significance of machine learning (ML) is emphasized in this paper to enhance existing system security. The review paper covers Denial of Service/Distributed Denial of Service (DoS/DDoS), cross-site scripting (XSS), SQL injection, and other network layer attacks. A total of 20 papers were compared based on the attack covered, performance, collected dataset, and the technique used in the proposed model.