E-Mail Spam Detection Using Refined MLP with Feature Selection

Abstract

—Electronic Mail (E-mail) has established a significant place in information user's life. E-Mails are used as a major and important mode of information sharing because emails are faster and effective way of communication. Email plays its important role of communication in both personal and professional aspects of one's life. The rapid increase in the number of account holders from last few decades and the increase in the volume of emails have generated various serious issues too. Emails are categorised into ham and spam emails. From past decades spam emails are spreading at a tremendous rate. These spam emails are illegitimate and unwanted emails that may contain junk, viruses, malicious codes, advertisements or threat messages to the authenticated account holders. This serious issue has generated a need for efficient and effective anti-spam filters that filter the email into spam or ham email. Spam filters prevent the spam emails from getting into user's inbox. Email spam filters can filter emails on content base or on header base. Various spam filters are labelled into two categories learning and non-machine learning techniques. This paper will discuss the process of filtering the emails into spam and ham using various techniques. Index Terms—Data Mining, Knowledge Discovery (KDD) Process, E-Mail, Spam, Ham, Spam Filter, N-Gram based feature selection, Multi-Layer Perceptron Neural Network (MLP-NN) and Support Vector Machine (SVM) classification algorithms.