Sentiment detection from Bangla text using contextual valency analysis

Abstract

Sentiment Analysis or opinion mining is an area of important research over the last decade. The basic task in sentiment analysis is classifying the polarity of a given text whether the expressed opinion in the text is positive, negative, or neutral. This paper presents an approach to sentiment assessment from Bangla text using contextual valence analysis. In linguistics valence of a verb is the number of satellite noun phrases with which a verb combines. We have used the WorldNet to get the senses of each word according to its parts of speech and SentiWordNet to get the prior valence (i.e. polarity) of each word. We calculate the total positivity, negativity and neutrality of sentence or document with respect to total sense. We developed our own methodology to calculate the sentiment from Bangla text using valency analysis. Sufficient examples and experiments are presented to describe the methodology.