Cyber Sentinel: Intelligent Phishing URL Identification System Employing Machine Learning Methods

Phishing attacks are a continuous threat to internet users, with the potential to cause identity theft, money losses, and other security lapses. The key to successfully preventing these threats is the automated detection of phishing URLs. This study presents a novel approach to detect phishing URLs that makes use of a range of Machine Learning (ML) techniques. Based on their inherent characteristics, the performance of multiple algorithms is assessed in between phishing and authentic URLs. K-Nearest Neighbors, Support Vector Machine, Random Forests, XGBoost, Decision Tree, Logistic Regression, and Multilayer Perceptron have been used. Using the suggested method, characteristics are taken from URLs, including lexical, domain-based, and content-based features including length of the URL, age of the domain, suspicious keyword presence, and resemblance to well-known phishing URLs. The impact of every algorithm’s is analyzed utilizing parameters such as accuracy, precision, recall, and F1-score on an extensive dataset that includes a sizable number of both phishing and genuine URLs. The proposed tests demonstrating the fact that XGBoost model has an impressive 97% accuracy rate in identifying phishing URLs, demonstrating the potency of the suggested strategy. Furthermore, the utilization methods for the use of Machine Learning in the in the real world to improve global cybersecurity measures by being integrated into web browsers, email filters, and network security solutions is discussed.