A hybrid approach to sentiment analysis

Abstract

This contribution presents a hybrid approach to Sentiment Analysis (SA) encompassing the use of semantic rules, fuzzy sets, unsupervised machine learning techniques and a sentiment lexicon improved with the support of Senti-WordNet. A Hybrid Standard Classification is first carried out, which is further enhanced into a Hybrid Advanced approach incorporating linguistic classification of semantic polarity modelled using fuzzy sets. The mechanism of the new SA methodology is illustrated by applying it to compute the polarity of a given sentence and to a benchmarking publicly available dataset: the Movie Review Dataset.