<![CDATA[Phishing is an increasingly prevalent cyber threat that is difficult to counter with traditional methods such as blacklists, which often fail to swiftly recognize new phishing websites. This research addresses the issue by developing a machine learning-based phishing URL detection system implemented as a browser extension called PhishBonk, using Logistic Regression for classification. The development process encompassed the collection of a dataset consisting of 134,850 legitimate URLs and 100,945 phishing URLs, data preprocessing, URL feature extraction, model training and evaluation, and finally integrating the trained Logistic Regression model into the PhishBonk extension to enable automatic real-time detection. Experimental results demonstrate that the Logistic Regression model achieved an accuracy of approximately 99.53% in distinguishing phishing URLs from legitimate ones. Furthermore, a System Usability Scale (SUS) evaluation yielded an average score of 81%, indicating that the PhishBonk extension is user-friendly and well-received. These findings suggest that the proposed machine learning-based browser extension effectively provides real-time, accurate phishing detection while ensuring a positive user experience.]]>
Copyrights © 2026
