Categorized classification of textual spam emails using data mining

Abstract

In the last few years, to boost the connectivity and the safety of the people in the world, the Internet has built various platforms. Across from them, Email is a substantial platform for people’s interaction. Email is an automated and efficient message transmission structure used to convey information from one person to another. And also, this appliance preserves plenty of money and time. Apart from that, Emails also have been abused by some assailants. The most commonly known one is Email Spam. Spam Emails are referred to as unwanted materials, and they are unrequested business Emails or forged Emails forwarded to specific personnel or sent to an Organization or spread among a set of people. For spam Email, users face many difficulties, such as increment of traffic, restricting the data processing duration and the storage area, consuming more time of users, and threatening user protection. Therefore, it is essential to have an appropriate Email filtering approach to secure Email. There has been plenty of general spam Email filtering systems and various research people's various efforts to classify Emails into ham (genuine Emails) or spam (fake Emails) using Machine Learning techniques. Unfortunately, most of the spam Email filtering solutions proposed so far are focused only on binary classification. However, classifying the already detected spam Emails into different types of categories is not performed yet. This research explored and proposed an effective system to categorize the Textual Spam Emails into different categories using Data mining. First, The Multi-Nominal Naïve Bayes classifier is applied to distinguish Emails into two groups, such as ham and spam. The Grouping algorithm is used to categorize the spam Emails, which are already obtained from the Naïve Bayes classifier, into distinct categories. (Finance, Health, Marketing, etc...)