Enhancing Network Security via Hybrid Intrusion Detection Systems with Machine Learning

International Journal of Computer Science and Technology | Vol. 11, No. 3, March 2009 | pp. 17–24

DOI: 10.46882/2009/IJCST/000012

Article Type: Research Article

Title: Enhancing Network Security via Hybrid Intrusion Detection Systems with Machine Learning

Names of Authors: Robert L. Vance¹, Sanjay Kumar², Elena Rostova³

Authors’ Affiliations: ¹Cybersecurity Research Center, Tech University, Austin, USA; ²Department of Information Technology, National Institute of Technology, New Delhi, India; ³Faculty of Computer Science, Saint Petersburg State University, St. Petersburg, Russia

Abstract: The proliferation of sophisticated cyber threats necessitates advanced security frameworks capable of identifying novel attack vectors. This study proposes a Hybrid Intrusion Detection System (HIDS) combining anomaly-based and signature-based detection mechanisms powered by a Support Vector Machine (SVM) and a Decision Tree classifier. By fusing these approaches, the system maintains high accuracy for known signatures while identifying zero-day variations. The model was evaluated using the benchmark KDD Cup 99 dataset, achieving an overall detection rate of 98.6% and a low false-positive rate of 0.42%. Experimental analysis indicates that the hybrid model outperforms standalone classifiers by a margin of 3.5% in multi-class classification tasks. Computational latency tests revealed an average processing time of 12.3 microseconds per packet, rendering it suitable for real-time intrusion screening in gigabit networks. The findings establish that integrating machine learning into hybrid setups significantly mitigates network vulnerability with minimal performance overhead.

Keywords: Intrusion Detection System, Hybrid Machine Learning, Support Vector Machine, Cyber Security, Packet Analysis, Network Anomaly

Manuscript Timeline: Received: October 5, 2008; Revised: December 2, 2008; Accepted: January 10, 2009; Published: February 12, 2009

Citation: Vance, R. L., Kumar, S., & Rostova, E. (2009). Enhancing Network Security via Hybrid Intrusion Detection Systems with Machine Learning. International Journal of Computer Science and Technology, 11(3), 17–24.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read