Bangla News Classification Using Machine Learning

Abstract

What is the distance between countries to countries from south to north, east to west in the earth? If you find the answer of this question according to the perspective of the present time, you will see that, actually, there is no distance at all. At present, people get the all sorts of news happening around the world instantly within couple of seconds. And it has become possible only because of virtual news portal. That is true that the online news portals are publishing news on live, but it is disappointing that users do not like all sorts of news published in the news portal. At that time, it has become need to a platform that can easily identify the user’s choice on news and publish only according to their choice. To classify the news by the user’s choice needs to analysis the news text. Lots of works has been done in English news by this time but there have very limited works on Bangla news. These make us inspired to do a research project on this topic though Bangla is one of the 8 major spoken languages around the world. In our research project, we deal with Bangla news collected from Prothom Alo newspaper. From preprocessing the news text, we try to do all sorts of procedures to classify the news text using Machine Learning classifier, “Naive Bayes classifier”. Finally, we develop a user interface to take the news text and show the class of that news.