Research Of Twice Semantic Retrieval Based On Conceptual Graph And WCN

As the main tool of information-obtaining in the Internet, search engine has been widely applied in all kinds of fields. Recent years, with more and more emergencies such as local war, terrorism, earthquakes, tsunami, snow disasters, fire disasters, infectious disease and serious accidents in coal safety production and electricity stable transportation, some crisis management, forewarn for emergency, model for emergency measures as well as the applied systems come into being at the opportune historic moment. However, they still provide users with search service of the model of "one size fit all", in which no difference of users in the design of sites is concerned and the same topological structure and way of presentation prevails on all visitors. Moreover, as a result of increasing of information and disorder of information arrangement in the Internet, they can't afford high-speed, precise, satisfied, various feedback for different users. This thesis poses a framework of twice semantic retrieval and also elaborates some key techniques concerned in accordance with mention above.Firstly, the thesis gives a general introduction of related knowledge about Conceptual Graphs, including comparison between Conceptual Graphs and other knowledge presentations, structures and formats of Conceptual Graphs, the mapping between Conceptual Graphs and Chinese, and tools in common use. Secondly, the thesis discusses the significance of user model in thesemantic retrieval framework and deals with the weighted conceptual network (WCN) in detail. The thesis also puts forwards represent model ofconception node storage structure of WCN, and also analyzes and discussesthe structual and regulative methods.Based on the discussion above, twice semantic retrieval framework is provided, which uses WCN as user model and expresses the real demands of users with the semantic relative relationship among the conceptions in Conceptual Graphs. Meanwhile, ways of storage of the Conceptual Graphs and WCN, the similarity algorithm between conceptual graph and WCN are provided. The algorithm is applied to rank the initial search result in second time which is of personality. The result of experiment indicates that Precision of information feedback has been raised efficiently.