Improving spam email detection using hybrid feature selection and sequential minimal optimisation

Abstract

Communication by email is counted as a popular manner through which users can exchange information. The email could be abused by spammers to spread suspicious content to the Internet users. Thus, the need to an effective way to detect spam emails are becoming clear to keep this information safe from malicious access. Many methods have been developed to address such a problem. In this paper, a machine learning technique is applied to detect spam emails. In this technique, a detection system based on sequential minimal optimization (SMO) is built to classify emails into two categories: spam and non-spam (ham). Each email is represented by a set of features extracted from its textual content. A hybrid feature selection is developed to choose a subset of these features based on their importance in process of the detection. This subset is then input into the SMO algorithm to make the detection decision. The use of such a technique provides an efficient protective mechanism to control spams. The experimental results show that the performance of the proposed method is promising compared with the existing methods.