Spam Text Detection in Bangla language and Comparison Between Machine Learning Algorithms

Abstract

Due to the advancement of technology electronic communication has reached its peak. It has great impact on our life. But some dishonest people are using it to fulfil their evil wish. Millions of people use Bangla text in internet. They communicate in e-mail, Facebook, twitter and other social networking platforms in Bangla language. And some people are manipulating this platform. They are sending spam messages to the users. Many works have been done for spam mail in English language. But there is no significant work for spam messages in Bangla language. This inspired us to work for spam message detection in Bangla language. Our approach to detect spam messages in Bangla language. For this approach 6 types of spam messages have been considered. We made the dataset for the task. We have collected the data from different platforms. Then we have used Natural Language processing to perform spam mail detection. We have compared between traditional NLP algorithm. For support vector machine we have got an accuracy of 96%.