The Research Of Fuzzy Domain Ontology Construction Based On Mining Online Reviews And Its Application

With the deepening of Internet services, Internet business services are well received by people, thus developing at fast speed. However, without the reference of the material objects, cyber citizens can not predict the quality of the goods. So they make their decision depending on the comments and idea of other consumers. Therefore, online comments become one of the most important standards of the goods on the internet. At present, the amounts of online comments increase at great speed. But the seller provide the comments of the goods through marking the score of the goods, result in the inadequate use of the abundant information included in the online comments. The topic of "How to use online comments to get full range of describing and evaluating the goods" has been the hot topic and emphasis on research.Ontology, as a shared concept of a particular area of the formal description is widely used in many areas. For the characteristics of universal emotion dictionary does not contain of dynamic emotional words and domain emotional words, we applied the ontology to the representation of knowledge of the online reviews. However, due to the classic ontology can not express fuzzy knowledge, and the emotional of vocabulary is uncertain. So in this paper, we applied the fuzzy domain ontology to express online reviews knowledge.The goods contain the characteristics of describing by domain emotional words and dynamic words, and online reviews contain comprehensive information on the commodities. We are discovering the combinations between properties and related domain emotional words and dynamic emotional words based on the associated word. In addition we use HowNet and the relationship between vocabularies to praise the two dimensions of emotional words. Finally, the semantic retrieval to maximize semantic match and sort results to the user.This article focuses on two aspects:(1) The area of fuzzy domain ontology construction: We design concept lattice based on weight generation algorithm for the associated knowledge mining, and compare with classical association rule mining algorithms Apriori to prove that the weights of concept lattice-based generation algorithm applies the high-dimensional sparse nature of the expression of text. During the fuzzy ontology construction process, we based on the framework of classical ontology, to use the annotations and assertions function of OWL2to describe fuzzy knowledge;(2) The area of fuzzy ontology application:Fuzzy domain ontology is applied to the weighted sentences emotional analysis. And through experiments to compare with emotional tendency analysis based on common emotional dictionary. We use accuracy, recall, and F values as the evaluation to prove that the effectiveness of the emotional tendency analysis based on the domain fuzzy ontology.