Research On Construction Of Automobile Domain Ontology Knowledge For Opinion Mining

With the rapid development of Web2.0, a large number of comments appears on the network, it has been a main aim at opinion mining how to capture the imply sentiment orientation from the subjective document, which contains the subjective opinion mining, attitude, feeling, emotion and so on. Currently, statistical methods are popular. It is uncertain in the topic, sentiment orientation, intensity of the polarity because the more complex grammatical structures in the sentence. It can not achieve the desired results.In this paper, we research the establish methods of ontology knowledge base for opinion mining from the real Web automobile reviews. We center on opinion mining of the Web automobile reviews, carry out the study on the construction of ontology system approach, ontology concept acquisition approach, ontology based feature opinion mining. The major works of this thesis include:(1) Construction of ontology system approachAccording to the concepts and the relation of automobile domain knowledge, the automobile ontology knowledge base for opinion mining is constructed. In this paper, we take the automobile parts, automobile performances, polarity dictionary to merge with ontology knowledge base. Constructed automobile ontology knowledge base for opinion mining can offer the powerful data resource to mine opinion whole information and special information of automobile. At the same time, it can improve the accuracy of polarity.(2) Ontology concept acquisition approachUsing the real Web automobile reviews corpus, the core concepts are automatic acquired in ontology knowledge base. In this paper, to adopt the methods of the template and bootstrapping can acquire the core concepts. It can provide conveniences for constructing the ontology knowledge base improve the efficiency of ontology construction and reduce the human resource. (3) Ontology based Feature Opinion MiningWe extract the features for opinion mining with the contracted automobile ontology knowledge base and related technologies, product a comprehensive summary by contacting opinion features with the the hierarchical relationship between the features. In this paper, by constructed the ontology we integrate the different representations of the same feature, classify the upper and lower features. Finally, we statistic the sentiment orientation of opinion feature, show them at different levels from top to bottom and generate a summary.