Design an Empirical Framework for Sentiment Analysis from Bangla Text using Machine Learning

Abstract

Natural Language Processing (NLP) lends a

helping hand for programming the computers to inspect a huge

amount of data. Sentiment analysis is an application of NLP

which deals with data to examine the sentiment or opinion that

can be either positive or negative. Using Bangla text, sentiment

analysis has become a challenge as there were only few works on

it. As a decision maker, sentiment extrication not only capturing

consumer attitudes but also helps in social behavior observance,

politics and policy making. This paper quantifies total positivity

and negativity against a document or sentence using Random

Forest Classifier to classify sentiments. We contemplate the use of

unigram, POS tagging, negation handling and classifier.