SQLI Attack: An Approach using ML and Hybrid Techniques

Abstract

Abstract: Web-based systems are significantly at risk from SQL Injection Attacks (SQLIA), particularly in industries that handle sensitive data, like finance and healthcare. During these attacks, hostile actors insert false SQL queries into the database server of a web application in an effort to steal sensitive data. The use of classifiers and techniques like end-to-end deep learning and expanding the Aho-Corasick algorithm to detect SQLIA attacks have been covered in the literature. These researches have shed light on identifying and minimizing SQLIA, but the problem still exists. To detect and prevent SQLIA, a thorough methodology is suggested that combines static and dynamic studies with machine learning and hybrid tactics. The study compares several machine learning algorithms with hybrid techniques, showcasing the hybrid strategy's higher performance in both training and test sets. This method, which provides a practical means to secure web applications, is put out as a potential remedy for the persistent problem of SQLIA. In result, SQL Injection Attacks continue to be a danger, requiring effective ways for detection and prevention. It is highlighted as a promising solution the potential of machine learning and hybrid solutions, notably the hybrid approach. The study emphasizes how crucial it is to follow recommended procedures for protecting online applications from SQLIA attacks.