Research On The Retrieval Model For Digital Library Based On Semantic Technology

With the development of network technology, the digital library has become an important means to obtain information. However, the digital library also has some disadvantages. The retrieval process are lack of semantic matching, only by matching keywords. Also, the information resources are lack of certain semantic. The emergence of the semantic provides the possibility to change the performance of the digital library retrieval.Dissertation introduces the problems of the traditional digital library and the study-situation of the ontology techonlogy. Based on these technologies,this dissertation puts forward a retrieval model for digital library. And implement a prototype system based on this model to check up the recall rate and the precision ratio.The main work of this dissertationis includes:1) To solve the problems of the digital library, the dissertaton proposed two kind of ontologies:the “Library Science” domain ontology and the metadata ontology. By analyzing the related books and the CLC number, determine the classes, individuals and so on. Finally build a domain ontology about the Library Science.Propose a metadata ontology of books based on DC metadata to solve the lack of certain semantic of the information resources.2) The dissertation put forward the process of query expansion based on the two ontologies. Use the SPARQL to match the user’s query, searching the upper class, the lower classes and individuals. Use the Jena reasoner to search the synonymous. Use the algorithm to sort these words. Finally use the Lucene tool to search the information in different dimensions.3) The dissertation construct a retrieval prototype using the above model structure. So that we can test and verify it.