Network Security Enhancement through Machine Learning–Driven Intrusion Detection

Abstract

With the rapid expansion of computer networks and internet-based services, protecting network infrastructures from cyberattacks has become a critical challenge. Traditional security mechanisms often fail to detect sophisticated and evolving intrusion patterns, highlighting the need for intelligent intrusion detection systems. This study addresses the problem of effective network intrusion detection by presenting a comparative analysis of instance-based and numerical machine learning techniques for a Network Intrusion Detection System (NIDS). In this work, K-Nearest Neighbor (KNN) and Support Vector Machine (SVM) models are employed to perform multiclass classification of network attacks, including Denial of Service (DoS), Probe, Remote-to-Local (R2L), and User-to-Root (U2R) attacks. The proposed framework analyzes network traffic patterns and behavioral features to identify malicious activities in real time. The models are trained and evaluated using benchmark intrusion datasets, and performance is assessed using metrics such as accuracy, precision, recall, and efficiency. Experimental results indicate that both KNN and SVM effectively detect multiple attack categories, with instance-based learning demonstrating strong detection capability. The study concludes that machine learning–based NIDS significantly enhances network security through improved accuracy, adaptability, and timely threat mitigation.