A Bangla Spam Email Detection and Datasets Creation Approach based on Machine Learning Algorithms
Ruhul Amin, Md. Moshiur Rahman, Nahid Hossain · 2019
Email is one of the most imperative communication mechanisms of the 21st century. People around the world send billions of emails every single day which makes people prone to threats. Spam emails can be used for stealing things from our electric devices, blackmailing and phishing. Moreover, we receive several unwanted emails such as advertisements and offers of e-commerce websites. Due to these emails, we sometimes miss important emails. There are several approaches available for detecting spam emails in the English language and other important languages. However, there is no such spam email detection tool available in Bangla language where Bangla is one of the most spoken languages in the world and widely used on the internet nowadays. For this reason, spammers nowadays send emails in Bangla language to Bengali people so that they can avoid being filtered out which makes us vulnerable to serious threats. Therefore, we have designed and developed a spam email detection mechanism in the Bangla language. We have also constructed datasets of Bangla spam emails to train and test our system. This paper explores the use of six supervised machine learning approaches. According to the classification results, Random Forest presented the best performance with 93.60% accuracy.