Spam/Ham Email Classification Using Machine Learning

Abstract

Email stands for Electronic mail which means it is the way to distribute messages by electronic means from a computer source to one or more by means of network. It is one of the fastest method of distributing messages between sender’s and receiver’s computer system via Internet. In these days, email has become the most frequent communication system and a significant portion of the population depends on accessible email or texts from strangers. One of the biggest issues email consumers confront is the increase of spam communications. The tools that determine if an email is spam or not are known as spam filters. Currently there are many spam filter tools available on internet. Identifying these spammers is one of the hot topic of research and arduous tasks. In this paper we will look into the spam filter techniques using machine learning. Logistic regression algorithm and the spam emails will be categorized using the Natural Language Toolkit. The dataset utilized comes from the public datasets of Apache Spam Assassin and includes samples of spam and ham. The accuracy is defined on the basis of score of cross validation, precession and recall.