Phishing email detection technique by using hybrid features

Abstract

Phishing emails is growing at an alarming rate in this few years. It has caused tremendous financial losses to internet users. Phishing techniques getting more advance everyday and this has created great challenge to the existing anti-phishing techniques. Hence, in this paper, we proposed to detect phishing emails through hybrids features. The hybrid features consist of content-based, URL-based, and behavior-based features. Based on a set of 500 phishing emails and 500 legitimate emails, the proposed method achieved overall accuracy of 97.25% and error rate of 2.75%. This promising result verify the effectiveness of the proposed hybrid features in detecting phishing email.