Hybrid approach for spam email detection

Abstract

On this era, email is a convenient way to enable the user to communicate everywhere in the world which it has the internet. It is because of the economic and fast method of communication. The email message can send to the single user or distribute to the group. Majority of the users does not know the life exclusive of e-mail. For this issue, it becomes an email as the medium of communication of a malicious person. This project aimed at Spam Email. This project concentrated on a hybrid approach namely Neural Network (NN) and Particle Swarm Optimization (PSO) designed to detect the spam emails. The comparisons between the hybrid approach for NN_PSO with GA algorithm and NN classifiers to show the best performance for spam detection. The Spambase used contains 1813 as spams (39.40%) and 2788 as non-spam (60.6%) implemented on these algorithms. The comparisons performance criteria based on accuracy, false positive, false negative, precision, recall and f-measure. The feature selection used by applying GA algorithm to reducing the redundant and irrelevant features. The performance of F-Measure shows that the hybrid NN_PSO, GA_NN and NN are 94.10%, 92.60% and 91.39% respectively. The results recommended using the hybrid of NN_PSO with GA algorithm for the best performance for spam email detection.