• Anjali Sahu,
• Dr. Ravindra Kumar Tiwari

The increasing reliance on TCP/IP-based web applications for financial transactions, healthcare, cloud computing, and enterprise communication has expanded the cyberattack surface. Conventional security mechanisms, including signature-based intrusion detection, static filtering, and web application firewalls, remain essential but may struggle with novel attacks, evolving traffic patterns, protocol anomalies, and legitimate-looking malicious requests. This paper proposes an integrated machine-learning framework for vulnerability detection and real-time threat mitigation. It combines network-flow analysis, HTTP characteristics, application security signals, supervised classification, anomaly detection, risk scoring, and policy-controlled responses. Random Forest and gradient-boosting classifiers provide computationally efficient baselines, while optional deep learning supports sequential traffic analysis. An adaptive response engine selects proportionate actions, including alerts, rate limiting, session challenges, temporary blocking, and human escalation. The evaluation methodology incorporates time-aware and group-aware data partitioning, class-imbalance handling, threshold calibration, and metrics such as precision, recall, F1-score, false-positive rate, detection latency, and mitigation effectiveness. Public intrusion-detection datasets can support preliminary network-level evaluation; dedicated web-application testing remains necessary to assess vulnerability-specific performance. The proposed framework jointly considers detection accuracy, operational latency, false-positive costs, and safe mitigation. Without experimental results, its effectiveness remains a testable research hypothesis rather than a demonstrated improvement.

Machine learning; web application security; TCP/IP; vulnerability detection; intrusion detection; threat mitigation; Random Forest; gradient boosting; anomaly detection; web application firewall; cybersecurity

"Machine Learning-Based Vulnerability Detection and Real-Time Threat Mitigation in TCP/IP-Based Web Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 10, page no.c332-c341, October-2026, Available :http://www.jetir.org/papers/JETIR2610239.pdf

"Machine Learning-Based Vulnerability Detection and Real-Time Threat Mitigation in TCP/IP-Based Web Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 10, page no. ppc332-c341, October-2026, Available at : http://www.jetir.org/papers/JETIR2610239.pdf