Research on the Sentiment Analysis of Customer Reviews Based on the Ontology of Phone

Abstract

In this paper, we develop FOBPRM (Feature Ontology Based Product Review Miner) system, to semi-automatically build the ontology tree of Phone area and extract the most representative expressions and customer opinions in the reviews, which represents for feature-sentiment pairs. Finally we develop our method of polarity calculation of feature-sentiment pairs and generate the all-round summary for customers and vendors. Instead of putting the emphasis on feature extraction and sentiment classification as the existing work did, we focus on the association between the features and sub-features of a product and their associated sentiment that influence the polarity of the attributes in fine-grained in this paper. Ontology built by computation of special degree and similarity degree has improved the accuracy and recall rate of features, and the information entropy computes the polarity of feature-sentiment pairs. The whole system works out the desired result.