COUNTERMEASURE TO MAN-IN-THE-MIDDLE ATTACK BASED ON EMAIL HIJACKING USING TRY-HYBRID SUPERVISED LEARNING TECHNIQUES

Abstract

Email communication faces an escalating threat from Man-in-the-Middle (MitM) attacks, which compromise the security and integrity of emails, leading to the risk of data breaches, financial losses, and reputational harm. Traditional email security measures, such as SSL/TLS encryption and authentication protocols (e.g., SPF, DKIM, DMARC), have become increasingly insufficient in countering these advanced MitM attacks. The growing sophistication of MitM techniques, including SSL stripping, DNS spoofing, and session hijacking. This research proposes a countermeasure to MitM attacks based on email hijacking using a try-hybrid supervised learning technique. timestamps, IP addresses, port numbers, packet sizes, and various security-related indicators. The development of the MitM attack detection technique employed a try-hybrid mitm attack detection technique, which combines the strengths of three machine learning algorithms: Random Forest, Gradient Boosting Machine (GBM), and Support Vector Machine (SVM).The results demonstrate the effectiveness of the proposed try-hybrid model, achieving an accuracy of 95.8%, surpassing Benchmark 1 (92.4%) and Benchmark 2 (90.1%). Precision improves to 94.3% compared to Benchmark 1 (91.0%) and Benchmark 2 (88.5%). Similarly, recall is enhanced to 96.5% against Benchmark 1 (89.7%) and Benchmark 2 (87.2%). The F1 score of 95.4% significantly outperforms Benchmark 1 (90.3%) and Benchmark 2 (87.8%). Moreover, the proposed model achieves a lower False Positive Rate (FPR) of 3.2% compared to Benchmark 1 (5.6%) and Benchmark 2 (6.8%).These results highlight the robustness and reliability of the try-hybrid model in enhancing email security by effectively detecting and mitigating advanced MitM attacks.