A Machine Learning-based Intelligent Web Application Firewall For Real-time Protection Against Sql Injection And Xss Attacks

Abstract

The Rapid Growth Of Web Applications Has Led To An Increased Attack Surface For Cyberattacks Such As Structured Query Language (SQL) Injection, Cross-Site Scripting (XSS), And Other Application-layer Exploits. Traditional Web Application Firewalls (WAFs) That Rely Solely On Static, Signature-based Rules Struggle To Detect Obfuscated Payloads, Zero-day Attacks, And Novel Variants Of Existing Threats. This Paper Proposes An Intelligent Hybrid WAF Architecture That Combines Signature-based, Anomaly-based, And Machine Learning–based Detection To Provide Robust, Real-time Protection For Modern Web Applications. The System Monitors And Filters Hypertext Transfer Protocol (HTTP) Traffic Between Clients And The Web Application, Using A Multi-stage Detection Engine To Identify Malicious Requests And Apply Appropriate Mitigation Actions. The Proposed Model Leverages Public And Synthetic Web Attack Datasets For Training And Evaluation, With A Focus On SQLi And XSS Detection While Remaining Extensible To Other Emerging Threats. Expected Outcomes Include Improved Detection Accuracy, Reduced False Positives And False Negatives, Scalability In Cloud-native Environments, And A User-friendly Monitoring Dashboard That Supports Effective Security Operations.

Keywords