Performance of Machine Learning Techniques for Email Spam Filtering

Abstract

Abstract-Email spam, an Unsolicited Commercial e-mail (UCE), is one of the major problems for today's Internet users. E-mail is an effective tool for communication as it saves time and cost, considerably. However, the increase in email users has resulted in a dramatic increase in spam emails during the past few years. There are several machine learning techniques such as Bayesian classification, Naive bayes, k-NN, SVMs, Rough sets etc., which help us with a technique to filter spam. In this paper, an email-filtering approach that is based on supervised classifier has been proposed.The model mentioned here has the advantage of combining supervised machine learning with Support Vector Machines (SVM).To implement SVM, we are using LIBSVM, a library from MATLAB tool and compared it with existing classifiers. The proposed solution for the undesired spam could achieve better results in the system function and is faster when compared to the other existing machine learning techniques.