A MULTI-CLASSIFIER PREDICTION MODEL FOR PHISHING DETECTION

Sarju S, Riju Thomas, Emilin Shyni C

Abstract: Phishing is the technique of stealing personal and sensitive information from an email client by sending emails that impersonates like the ones from some trustworthy organizations. Phishing mails are a specific type of spam mails; however the effects of them are much more terrible than alternate sorts. Mostly the phishing attackers aim the clients of the financial organizations, so its detection needs high priority. Lots of research activities are done to detect the phished emails, in the proposed methodology a multi-classifier prediction model is introduced for detecting phished emails. Our contention is that solitary classifier prediction might not be satisfactory to urge the clearest picture of the phishing email; multi-classifier prediction has accuracy 99.8% with an FP rate of 0.8%.

Keywords: Phishing email detection, Machine learning techniques, Multi-classifier prediction model, Majority voting

DOI: https://doi.org/10.15623/ijret.2014.0303005