Ontology Based Method for Sentiment Examination

Abstract

In recent years, sentiment analysis became an important concept which is used to extract an important feature or text mining according to our interest. It is very difficult to judge the polarity or the review of particular product. This problem is solved by ontology. Ontology is a formal and shared specification of a particular domain. It is a large semantic network used to extract a domain specific concept or domain specific features of a particular product. Ontology contains concept which means domain and sets of objects or we can say instance of concept and relationship between various objects. Ontology based approach is classification of opinion and feature-based approach. Ontology gives domain specific features along with the hieratical relationship between them. Ontology generates semantic graphs and tree for showing relationship between different domains and entities. The relationship can be done on the basis of nodes and edges connections. We use some contextual lexicons which are used to provide the opinion about particular domain like SentiWordNet, SenticNet, WordNet etc. This survey aims to provide an insight on Opinion target, Opinion lexical and Aspect based Polarity detection, sentiment aggregation.