| With the flourish development of network and communication technology, we accumulate a mass of data information, which becomes indispensable to our lives. These data have obvious distributed, autonomous and heterogeneous property. Most search engine and some integrated software treat the query of these heterogeneous data often cannot play biggest role presently. The result also often has only "scope" but no "depth" and exists plenty of rubbishes. Query mechanism is inflexible and hasn't point. The personalized query of heterogeneous data has become the bottleneck problem of data sharing.A kind of personalized query platform is researched and implemented in this dissertation. In order to assure the recall and accuracy of query and realize personalized query, in the research course of this platform, Agent and Ontology technology are led into.As a domain model, Ontology provides a common comprehension about the domain concepts and the hierarchy of the concepts; at the same time, its utilization also reduces the dependence on the comprehension technology of natural language. The main contributions of this dissertation are listed as below:· An Ontology-based user model is given. A user model in a special information domain is a set of personalized domain Ontology and personalized information need, which contain the user's information.· An Ontology-based algorithm of user modeling is given. User modeling is a dynamic process during user browsing and retrieving information. The computing of the interest changes of user to the concepts is the stress of the algorithm.· The prototype system is complemented. It can monitor the process of user's browsing the information real time, record the content which user has visited and user's feedback, then analyze and deal with these datas, build user's personalized mode, It provides an interactive interface, by which user can learn and build his orher personalized information requirement to implement personalized retrievalservice.
|
PDF Full Text Request