Ham and Spam E-Mails Classification Using Machine Learning Techniques

Abstract

Spam e-mail has become a very serious problem. Sending inappropriate messages to a large number of recipients indiscriminately has resulted in anger by users but large profits for spammers. This article looks at classifying spam e-mails from inboxes. Ten alternative classifiers are applied on one benchmark dataset to evaluate which classifier gives better result. A 10-fold cross validation is used to provide the accuracy. Results of the classification algorithms are compared with the spambase UCI dataset. The experimental results approve that the spam mails can be classified correctly, with accuracy reaching up to 95.45% for the Random Forest technique, compared to other classifiers used.