Machine Learning Based Intrusion Detection System for Network Security

Mr. Kishor Golla¹, Thoom Shree Mani Rao²

¹Professor, Department of Computer Science and Engineering, St. Martin’s Engineering College, Hyderabad, India kishorgolla1984@gmail.com

2Student, Department of Computer Science and Engineering, St. Martin’s Engineering College, Hyderabad, India shreemanirao@gmail.com

Abstract:The rapid expansion of internet-connected devices, the transition to 5G networks, and the proliferation of cloud computing have exponentially increased the attack surface for malicious actors. Traditional network security mechanisms, primarily signature-based Intrusion Detection Systems (IDS), are increasingly inadequate against zero-day vulnerabilities, polymorphic malware, and sophisticated Advanced Persistent Threats (APTs). These conventional systems rely on pre-existing databases of known attack patterns, rendering them blind to novel anomalies. This research proposes a highly robust, anomaly-based IDS leveraging supervised Machine Learning (ML) algorithms to intelligently classify network traffic. By utilizing the benchmark NSL-KDD dataset—a refined iteration of the KDD CUP 99 dataset—this study implements a comprehensive data preprocessing pipeline, including feature selection via Information Gain, to reduce computational overhead. We evaluate the performance of various classification models, specifically focusing on Random Forest (RF) and Support Vector Machines (SVM). Experimental results demonstrate that the Random Forest ensemble model achieves a superior detection accuracy of 98.2%, a precision rate of 97.5%, and significantly minimizes the False Positive Rate (FPR) compared to traditional algorithms. This research establishes that ML-driven anomaly detection provides a scalable, highly accurate defense mechanism for modern, high-speed network architectures. Keywords: Intrusion Detection System (IDS), Machine Learning, Anomaly Detection, Network Security, NSL-KDD Dataset, Random Forest, Feature Engineering, Support Vector Machine (SVM), Cybersecurity.