Sentiment mining: An approach for Bengali and Tamil tweets

Abstract

This paper presents a proposed work for extracting the sentiments from tweets in Indian Language. We proposed a system that deal with the goal to extract the sentiments from Bengali & Tamil tweets. Our aim is to classify a given Bengali or Tamil tweets into three sentiment classes namely positive, negative or neutral. In recent time, Twitter gain much attention to NLP researchers as it is most widely used platform that allows the user to share there opinion in form of tweets. The proposed methodology used unigram and bi-gram models along with different supervised machine learning techniques. We also consider the use of features generated from lexical resources such as Wordnets and Emoticons Tagger.