Spam mail detection using data mining classifiers

Abstract

Introduction: Kumar, Poonkuzhali and Sudhakar (2012) state that email is the quickest andmost cost-effective medium of communication, allowing internet users to eas-ily transfer information from anywhere in the world in a fraction of second. Rathi and Pareek (2013) and Basavaraju and Prabhakar (2010) state that ifthe email address and sender’s identity are both anonymous, and the communication is sent to a group of receivers who are not interested in receiving it, then the email is characterised as spam mail. According to Basavaraju andPrabhakar (2010), email is subject to spam because of its widespread use andlow cost. According to Kumaret al. (2012) “telemarketers are able to access personal information from service providers and at times service providers sell the information to telemarketers. Hence a user’s personal information could get accessed by others without the users permission. Telemarketers use the personal information to send bulk emails to recipients without gaining their permissionto do so”. As a result, the recipient’s mailbox gets bombarded with emails,r esulting in spam email. Significance of the study: The importance of spam mail detection has become more essential than ever (Kiwanuka, Alqatawna, Amin, Paul and Faris, 2019). According to Cook, Manderson, Hartnett and Scanlan (2005), email spam is a big problem that causes financial harm to businesses and negatively impacts email users. Manually deleting spam mails from an inbox is inefficient and if the user is atwork then it is also costing their employer money in lost productivity. Spam mail has also been known to cause network outages. Kelin (2001) states that money spent on internet access, time spent downloading, reading and deleting the spam are examples of expenses imposed on the recipient...