A New Ontology-Based Method for Arabic Sentiment Analysis

Arabic sentiment analysis is a process that aims to extract the subjective opinions of different users about subjects since these and sentiments are used recognize their perspectives judgments in particular domain. Few research studies addressed semantic-oriented approaches for based on domain ontologies features’ importance. In this paper, we built semantic orientation approach calculating overall polarity from texts ontology available lexicon. We concepts weight features by considering levels tree frequencies dataset compute given textual review importance each feature. For evaluation, an hotels’ was selected build test proposed approach. The accuracy f-measure reach 79.20% 78.75%, respectively. Results showed outperformed other approaches, it appealing be analysis.