URL Based Detection of Phishing Using Random Forest Algorithm and SVM
Abstract
Malicious sites generally advance the development of Internet crimes and oblige the improvement of web services. As a result, there has been a strong inspiration to make basic responses to keep the customer from visiting such destinations by making the customer aware of the threat the site presents. We propose a learning based approach to manage request destinations into 3 classes: Benign, Malware and Malicious. Our mechanism just separates the Uniform Resource Locator itself without accessing the contents of the site pages. Thus, it kills the run-time latency and the possibility of introducing customers to the browser based vulnerabilities. This paper proposes a feasible course for the area of phishing destinations using Random Forest Algorithm, SVM and few specific URL features for better results.