An opinion analysis model for implicit aspect expressions based on semantic ontology

Abstract

Analyzing user opinion and building features model of entities is very important for sentiment analysis and opinion mining, which is the fundamental work for social network analysis. Opinions of entity can be classified into two types: explicit aspect expressions and implicit aspect impressions, and implicit aspect expressions might have some relationships which can be denoted by domain specific knowledge, existing previous works have not addressed in much detail. This paper defines an opinion mining model for implicit aspect expression based on semantic ontology, and proposes semantic similarity estimation approaches of entity by the domain ontology.