Filter Spamming In Computer Networks by Text Mining and Machine Learning Method

Abstract

The goal and motivation of spam is transmitting information to the recipient, such information including an advertisement for a product (which is likely to be worthless, illegal or non-existent), prey to a fraudulent scheme or to promote a cause or destroy a computer, which often designate to influence or stealing the recipient's computer. Due to costless of sending information via an e-mail, a method is used for spammers and those who have specific targets. Only a very small part of the recipients may be as a target. This number could be one in ten thousand or less. But prevent and filtering the spam could improve the network traffic. Although in recent years the introduction of a large number of text classification algorithms have been proposed, but text mining requires a series of pre-processing, effective, efficient and fast learner algorithms. In this paper, a series of experiments are performed on spam data, and a method with high reliability and low error rates have been provided