The Design And Implementation Of Hownet-Based Semantic Retrieval Model

Web is heterogeneous, dynamic, huge and distributed. The properties of Web make the retrieval of web data difficult, for the data is only fit for human to read rather than search engines. The most common search engines are on a syntactic base which failed to reach high precision and recall. As the account of web pages growing, this kind of search engine can not satisfy people's requirement of obtaining knowledge rapidly.To improve this situation, people made a lot of efforts to research advanced technologies in artificial intelligence and natural language processing. However, more and more complicated technologies are of poor help to resolve the essential problem of information retrieval. Semantic Web was raised in this context. It changes the expression of web data and makes it more clear and formulized, which can be used for search engine to understand web users' intention, to express documents and media, and then to match them in a semantic level. However, semantic retrieval using semantic web technologies are still faced with difficulties such as ontology acquisition, ontology integration, etc.To solve such problems, this research proposes an approach to convert HowNet to ontology in OWL language. And then a fresh semantic retrieval model based on this HowNet ontology was raised. Furthermore, a prototype system based on the proposed model is developed as well as verified and analyzed by comparing with other semantic systems.