Feature Fusion and Clustering for Improved Intrusion Detection Accuracy: A Comparative Study

In the presence of cyber threats that are becoming more sophisticated, network security continues to be a critical concern. The necessity for more sophisticated detection methods is frequently prompted by the inability of conventional intrusion detection systems to keep pace. We present a comprehensive approach to network intrusion detection in this study that utilises machine learning techniques, feature selection, and clustering. These systems are essential for safeguarding sensitive information and guaranteeing the security of digital infrastructure. A remarkable level of accuracy was demonstrated in the experimental results by the use of feature selection, ensemble learning, clustering, and a variety of analysis methods. The system’s high efficacy in detecting intrusions is demonstrated by the application of key performance metrics, including accuracy, detection rate, precision, F1-score, and Cohen’s Kappa to evaluate the proposed framework. This work demonstrates the potential of integrating multiple machine learning models and clustering techniques to enhance detection accuracy and reduce false positives, thereby providing a powerful solution for bolstering network security against contemporary cyber threats.