Phishing Website Detection Using Random Forest and Support Vector Machine: A Comparison
Norzaidah Binti Md. Noh, Megawati Basri
Abstract
As the Internet usage is growing rapidly, people are changing their choice from traditional shopping to electronic commerce. However, instead of robbery at banks and stores, criminals now locate their victims with some tricks in cyber world by applying the anonymous internet framework. Hackers are using new tactics, including phishing, to mislead the victims by using fake websites to gather sensitive information, including account ids, usernames, and password. Understanding whether a website is legitimate or phishing is an incredibly difficult problem because of its phishing attack structure, which primarily targets the vulnerabilities of web users. This work was proposed to solve these issues by using machine learning technology to detect the phishing websites. The system analyzed the HTML code structure that include in the hyperlink of the websites. Two machine learning techniques, namely Random Forest and Support Vector Machine are tested to identify the best machine learning algorithm in detecting phishing website. The performance metric for each algorithm was used as measured. The result average of accuracy for Random Forest is 99.98 percent and Support Vector Machine is 84.73 percent. This study aims to detect phishing websites hyperlink and produces the best algorithm which is Random Forest that achieved the highest accuracy to be used for the system.