A Hybrid Approach for Identifying and Classifying Darknet Traffic Patterns
K. G. Anilkumar, D. Muhammad Noorul Mubarak · 2024
In today’s digital landscape, the proliferation of digital devices has created a pressing need to safeguard the sensitive information they store. As a result, researchers are intensifying their efforts to develop more robust security measures. However, the rapid evolution of machine learning algorithms has created a demand for even more effective and efficient threat detection models. To address this challenge, we have developed the Enhanced Smell Detection Algorithm (ESDA), a novel feature selection technique that, when combined with the XGBoost classifier, significantly improves the detection of darknet-related cyber threats. By selecting the most relevant features, ESDA reduces computational complexity, while the XGBoost classifier’s iterative approach ensures consistent and superior performance in identifying and mitigating darknet-borne threats.