MACHINE LEARNING-BASED PHISHING URL DETECTION USING LOGIN URL PATTERNS

Abstract

This paper recommends a machine learning-based approach that employs login URL patterns to identify phishing URLs in order to enhance the accuracy and efficiency of the process of identifying harmful web links. Looking at the structural and lexical characteristics of URLs, particularly those associated with login pages, is the recommended approach. Phishing attempts frequently employ questionable subdomains, misleading character patterns, and unusual domain names. The system efficiently detects hidden patterns and categorizes them by utilizing supervised learning algorithms and a labeled dataset of legitimate and malicious URLs. Feature extraction techniques improve detection performance, even against unknown threats, by capturing both syntactic and behavioral characteristics. The experimental results demonstrate that the model outperforms traditional blacklist and rule-based approaches in terms of precision, recall, and accuracy. This research examines the potential of machine learning to enhance cybersecurity by providing a more adaptable and scalable approach to combating the rapid evolution of phishing schemes.