Predictive data mining for phishing websites: a rule based approach

Abstract

The rapid advancement in internet has paved way for several serious crimes, of which phishing occupies a very important place. Phishing is a form of cybercrime where an attacker mimicking a legitimate website or a person or an organization redirects the victims to steal confidential data through e-mail, malwares or some other social engineering platforms. Victims prominently suffer from financial loss and private data loss. The serious outbreak of phishing has paved way for many researches, though comprehensive and accurate solution has not been proposed so far for thwarting its impact. This paper aims to develop a resilient model to predict phishing scam by means of classification algorithms of data mining. Five algorithms were chosen for this purpose and a comparative study was undertaken for their performances, accuracy, error rate and efficiency. The rules generated from the algorithms showed up a relatively better performance than the existing phishing detection tools