FakeURLGuard: Real-Time Web Security Enhancement Using Machine Learning-Based Fake URL Detection and Prevention Techniques

Phishing denotes to an attempt to steal profound information. Phishing URLs attacks remain one of the most deceptive URLs to harvest sensitive user data. This paper presents FakeURLGuard system that provide a real-time web security for the web users. It includes lexical and statistical features which directly extracted from the URL string. Our FakeURLGuard system does not require page content or network access. FakeURLGuard system utilizes three supervised machine learning classifiers with soft voting ensemble techniques. Our experimental results show that the voting classifier attained the highest correctness of 97.55%, outperforming each individual base classifier. Our dataset containing 10,000 URLs with 5000 URL phishing and 5000 URL benign.  The proposed FakeURLGuard system deployed using Flask framework. It enabling phishing URL detection and prevention through an interactive user interface in real-time.