Phishing Email Detection Based on Hybrid Features

Abstract

As an attack of social engineering, phishing email has caused tremendous financial loss to recipients. Therefore, there is an urgent need for phishing email detection with high accuracy. In this paper, we proposed phishing emails detection based on hybrid features. By analysing the email-header structure, email-URL information, email-script function and email psychological features, we extracted 18 hybrid features. Then we chose Support Vector Machine (SVM) classifier to evaluate our experiments. Experiments are performed on a dataset consisting of 500 legitimate emails and 500 phishing emails. The proposed approach achieved overall true-positive rate of 99%, false-positive rate of 9%, precision of 91.7% and accuracy of 95.00%. Furthermore, we evaluated the effectiveness of our proposed psychological features. The results showed that psychological features can improve the accuracy of detection and reduce the false-positive rate. Our proposed method has a good performance in detecting phishing emails.