Based On The Semantic Web Information Retrieval Technology

At the XML2000 Conference in December 2000, Tim Berners-Lee proposed a semantic web, the concept of the next generation of Internet, in an effort to assign an unique identity to all resources over the WAN, and to build semantic associations amongst all resources that can be processed by machine. Extending the current WAN concept, the semantic web uses a distinctive and formal representation for information as an effective means for indexing and accessing between heterogeneous systems, thus implementing the interconnection of information resources in terms of semantics. With such an interconnection, it is possible to implement high-level intelligent applications based knowledge for the purpose of knowledge sharing.At present, the Internet has major flaws in terms of information representation and retrieval, since it is designed for users to directly read and process information and it does not provide the semantic information for computers to understand, thus limiting their capability in automatically analyzing information retrieval and intelligently processing information. Therefore, a great challenge in information retrieval is how to implement the semantic retrieval on information resources to make good use of the available digital resources.This thesis makes an effort to build a framework for information retrieval over the WAN based on the semantic web, describe the design methodology and search flow for information retrieval, and state the rational behind the system model approach. This work also includes an overview on the basic principles, techniques, available tools, and the recent advancements in information retrieval.After an in-depth study on key techniques associated with model building in intelligent information retrieval, this work proposes an effective approach toward such models that can employed as the base to construct an experimental system for intelligent retrieval. These key techniques include domain ontology structure, information resource collection, semantic inference, and ranking of retrieval results. Furthermore, based on theoretical analysis, an experimental system for literature retrieval based on the semantic web is designed for computer literature retrieval that builds the literature ontology and the semantic dictionary proper as the domain ontology. Through a mapping from literature resources to knowledge level and a semantic inference, the experimental system under study can mine (discover) the semantic associations amongst different literatures, thus offering the needed semantic information to the targeted resource in literature retrieval, and making semantic retrieval possible for users. The experimental system implements a relatively complex knowledge retrieval and a second-level retrieval, as well as the valued added service for domain resources. These functionalities can hardly be implemented by the traditional key word-based retrieval methods, and through experiments the study has verified the feasibility of a system model for information retrieval.