INTERNATIONAL JOURNAL OF ADVANCE RESEARCH
IN MULTIDISCIPLINARY

Phishing scams have increasingly become a significant cybersecurity concern, luring users into disclosing sensitive information through fake websites. Enter PhishShield, an innovative phishing detection system that utilizes machine learning, seamlessly integrated into a browser extension. At its core, PhishShield harnesses a trained Random Forest model to analyze URLs and classify them as either phishing attempts or legitimate sites based on a variety of extracted features. The backend, developed using Python and Flask, hosts the model and serves as the API for processing URL queries. On the other hand, the frontend operates as a Chrome extension that delivers real-time detection capabilities. With PhishShield, users can navigate the internet with greater confidence, as the tool actively works to prevent them from becoming victims of phishing attacks. To achieve accurate classifications, PhishShield employs a comprehensive Random Forest classifier trained on a vast dataset of both phishing and legitimate URLs. It extracts a multitude of features related to the URLs, such as their length, the inclusion of special characters, the use of IP addresses, domain age, and SSL certificate validity. These specially designed features allow the model to effectively pinpoint deceptive websites, steering clear of traditional blacklist methods. The backend effectively manages feature extraction and model inference, responding promptly to requests from the frontend. When a user visits a page, the Chrome extension captures the current URL and sends it for classification. If the site turns out to be suspicious, users receive an immediate warning, advising them not to proceed, thus providing an extra layer of security in their online activities.