Polarity Detection at Sentence Level

Abstract

Opinions bear a very important place in the life of human beings. Human Beings are always surrounded by opinions, when a decision has to be taken; people always want to know the opinions of others. But as the impact of the web is increasing day by day, Web documents can be seen as a new source of opinions for the people. Large numbers of reviews are available on the Web related to every product. Whenever a customer buys any product, they express their feedbacks as opinions on the e-commerce website, thus it is very important to automatically analyze the huge amount of information on the web and develop methods to automatically classify the reviews. Opinion Mining or Sentiment Analysis is the mining of attitudes, opinions, and emotions automatically from text, speech, and database sources through Natural Language Processing (NLP). In this paper an opinion mining system is proposed using unsupervised technique to determine the polarity of sentences i.e. to classify the sentences as positive, negative or neutral. Negation is also handled in the proposed system. Experimental results using reviews of products show the effectiveness of the system.