Study On Information Retrieval Model Guided With Query Conceptual Graph

Information Retrieval (IR) Model has been developed for more than70years. In a long period of time, only the trained people can use it andthe model was simply implemented, common people could acquireinformation by many other ways, so that they were not urgent forinformation Retrieval. But it was growing up gradually with the rise of theInternet. In such a numerous and complex ocean of information, how toaccurately acquire the satisfactory ones become a highly required task inIR research. Now many search engines give service to everyone, but insome aspects the performance are not encouraging. The main reason is thatthey chop the user query into some keywords which are isolated each otherand it doesn’t keep the query as an entirety concept, therefore the semanticinformation between keywords is lost.We firstly start from the theory of conceptual graph research, stressthe importance of conceptual analysis on the characterization of user queryintent. The IR model based on conceptual graph can improve the precisionof retrieval by using query conceptual graph to maintain the query’ssemantic meaning, integrating the conceptual graph matching and semanticsimilarity calculation. Then we give an objective analysis of the difficultiesto implement such model, and reconsider the relevant technology, at lastour innovation is in proposing the IR model guided with query conceptualgraph.Around this model, this paper first focuses on query conceptual graphindexing, and analyses the relevant characteristics of user query. Combinedwith the some methods like vocabulary knowledge acquisition, we discussthe method to complete the indexing of the <E-A-V> form of conceptual graph. Secondly, we give a detail introduction of the semantic similaritycalculation and the conceptual graph similarity calculation and show thecomparison of the advantages and disadvantages of each. Finally we statethe details of implementation, introduce practical experience in theapplication of the conceptual graph, and provide some useful ideas for thefuture conceptual graph completely used in information retrieval andsemantic search.