Ontological Filtering for Sentiment Analysis

The rapid growth of the internet has increased the number of online reviews, opinions and sentiments toward products, services or topics. People appreciate this opportunity so that e-Commerce websites provide services for users to publish their reviews. Social networks, blogs and websites enable, thanks to the reviews, a social structure that provides benefits for the users and the firms that hosts electronic markets. Therefore, this huge quantity of information can confuse users and does not produce useful knowledge. In such a context, in fact, who says what and how they say it, matters. In this scenario a valuable contribute can be given by the sentiment analysis that is one of the hottest current research area. This paper presents a novel approach to the sentiment analysis which is based on the ontological filtering approach. The proposed approach shows how to automatically mine, from a corpus of documents, positive and negative sentiments .Experimental evaluations, on real dataset, show that the proposed approach is effective and furnishes interesting results.