Sentiment Analysis of Bangla Text Using Deep Learning

Abstract

Sentiment analysis is a set of instructions to define and derive opinions. Most of the Sentiment Analysis works centered on the English language. Bangla is widely spoken language, approximately 265 million of people uses this language for communication. In pursuance of English language, the huge and standard datasets are not found for Bangla Language. For the lack of standard datasets, the research works for Bangla has been shorten in pursuance of other languages. To recognize the sentiment from Bangla texts is the main target of this research work. In this research work, we have tried to build such a system, which will recognize the sentiments from Bangla texts. It is very challenging to work on such a language which does not have enough standard data. In the end, we have successfully got the sentiments from the Bangla texts. The datasets were collected from social medias and Bangla blogs. Data were collected and trained the data to the system. Lastly, the system shows the expected result which presented into the research. Working with Bangla text is a quite challenging job, there are not enough resources or data sets available in Bangla. Collecting data was a huge challenge for us. We try to collect data from different sources like newspapers, blogs, some social media comments, and status.