Corpus Generation for Multi-class Bangla Email Classification and Its Evaluation
Md. Tariqul Islam, K. M. Azharul Hasan · 2023
Email is an important communication tool used worldwide, but it also exposes people to various threats such as spam emails, which can result in theft, blackmailing, and phishing. Unfortunately, there is no spam email detection tool available for the Bangla language, despite it being widely used on the internet and email This makes Bengali people are vulnerable to serious threats as spammers send emails in Bangla to avoid filters. There is a huge number of emails sent in Bangla and these Bangla emails suffer from such problems. Though three is no dataset available for Multi-class Bangla Email Classiflcation, T'his paper explores generating a corpus for Bangla emails using important features that help to filter out Bangla emails into multi-classes. We generate a collection of 7000 Bangla emails to address the problem. We also evaluate the features selected of the emails for multi class classification.