Email Spam Detection by Machine Learning Approaches: A Review

Currently, technology has exhibited substantial advancement, resulting in the improvement of communication. Emails are often regarded as the most effective method for both informal and formal communication. Furthermore, individuals utilize email as a means to save and distribute significant data, encompassing textual content, images, documents, and various other things. Due to emails’ simple and easy-to-use nature, some people abuse this mode of communication by sending an excessive amount of unwanted emails, usually referred to as spam emails. The spam emails may include malicious content that is disguised as attachments or URLs, posing a risk of security breaches to the host system and potential theft of sensitive information such as credit card data. These days, spam detection poses serious and massive challenges to email and IoT service providers. Various previous studies have concentrated on machine-learning methods to detect spam emails in the mailbox. The primary aim of this work is to provide a comprehensive examination and comparative evaluation of machine learning techniques utilized in the detection of email spam. Also, it highlights the main challenges that face spam email detection. Furthermore, a thorough evaluation of various strategies is conducted, taking into account metrics such as accuracy, precision, recall, and F1-score. Finally, a thorough analysis and potential areas for future research are also examined.