Year: 2025
Type: article
Abstract: The dynamic nature of modern cyberattacks has placed significant pressure on traditional defense mechanisms, which often fail to adapt against novel threats.As network infrastructures expand and the volume of data traffic grows, the role of intelligent Network Intrusion Detection Systems (NIDS) has become crucial in identifying and mitigating malicious activities before they compromise security.This research introduces a machine learningdriven NIDS that employs URL-based feature analysis to detect potential intrusions within network traffic.Unlike conventional rule-based systems, the proposed approach focuses on extracting lexical, domain-specific, and statistical characteristics from URLs, enabling the model to capture subtle patterns that distinguish benign from harmful connections.A diverse dataset comprising legitimate and phishing URLs was processed to develop a comprehensive feature set.Several classification algorithms were tested, including Decision Trees, Random Forests, and Support Vector Machines, with Gradient Boosting emerging as the most effective model.Its ensemble learning capability provided strong generalization, reduced bias, and improved resilience to imbalanced data.The experimental evaluation demonstrated that the system not only achieved high accuracy but also maintained a low falsepositive rate, addressing one of the persistent challenges in intrusion detection..
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W3217222811