Phishing Websites Prediction Using Classification Techniques
Dyana Rashid Ibrahim, Ali Hussein Hadi · 2017
Phishing is an important issue that faces the cyber security. This paper exploits the capabilities of classification techniques on Phishing Website Prediction (PWP), and introduces a methodology to protect users from the attackers. The blacklist procedure isn't a strong enough way to stay safe from the cybercriminals. Therefore, phishing website indicators have to be considered for this purpose, with the existence and usage of machine learning algorithms. Five different classification techniques have been used to evaluate their efficiency on (PWP) in terms of accuracy and the Relative Absolute Error (RAE) value for each one of them, with and without the feature selection process. WEKA tool was used for the implementation of these classifiers on a public dataset from NASA repository. The motivation behind this investigation is to employ a number of Data Mining (DM) algorithms for the prediction purpose of phishing websites and compare their effectiveness in terms of accuracy and RAE. Where DM classifiers have proved their goodness in this kind of problems.