Publication

Abstract : Web applications are essential components of today’s digital ecosystems, allowing for smooth interactions and data transmission. The persistent concerns of SQL injection (SQLi) and Cross-Site Scripting (XSS) attacks, on the other hand, continue to pose substantial problems to the security of these systems. Traditional security methods are frequently ineffective in mitigating these risks, demanding a more dynamic and adaptive approach. This research investigates a novel way to bolster web application security by employing machine learning, specifically the k-Nearest Neighbors (KNN) algorithm, for attack payload generation. The implementation of the k-Nearest Neighbors (KNN) technique for creating malicious payloads in the context of SQL injection (SQLi) and Cross-Site Scripting (XSS) attacks on web applications is investigated in this work. The work presents a prototype system that incorporates KNN into the payload generation process, proving its effectiveness in constructing context-aware payloads that elude standard detection systems. The results of the prototype system demonstrate the method’s usefulness in circumventing typical security protections. This study provides valuable insights into improving web application security by emphasizing the role of machine learning in proactively tackling growing cybersecurity threats.