CLUSTERING OF EMAILS USING ANT COLONY ALGORITHM

Abstract

Organizations and individuals depend heavily on email system as one of the major sources of communication. Electronic mail (email) has become one of the most powerful communication tools for both individuals and organizations. The number of emails received daily increases sometimes resulting to the problem of email overload which can become a burden to email users. Therefore the need for a clustering or classification system that will categorize these email data into different groups in other to enable users manage their email box efficiently. Data mining and analysis can be conducted for several purposes such as: clustering and classification, Spam detection, subject classification and so on. Different techniques have achieved a significant performance in email and document clustering in the past. This research deal with email clustering using Ant Colony Algorithm (ACO). A model based on ACO was developed which consists of two modules, email documents pre-processing and automatic clustering module. A large set of newsgroup email data set was used. The research model was implemented using MATLAB. The performance evaluation of the model shows that it has high significant performance in terms of clustering accuracy, when compared with some existing models.