Researeh And Implementatio On Personalized Information Service For Digital Library Based On Ontology

The rapid development of Internet provides the basic condition for thedevelopment of digital library. But it also brings the trouble of information explosion,although the information on digital library is relatively in order. All kinds of informationexisting in different databases in digital library loads our users with too great a burden.We utilize the chance of the research on personalized information server for the libraryof Zhejiang Industry Polytechnic College, the project titled as “Research andImplementation on Personalized Information Service for Digital Library Based onOntology” is launched. We import Ontology in the project and carry out the preliminaryexploration.The system of personalized information service for digital library base on ontologyincludes user layer, resource layer and business layer. The concrete works include theconstruct of ontology database, the user interest model based on context informationand the similarity algorithm between concepts.(1)We Launch user requirement analysis and construct the system frame from therequirement of resource database, the requirement of user service and the business, sothe system is formed by the user layer, the business layer and resource layer.(2)We propose the rules of ontology database based on subject thesaurus. Thedomain ontology database of education discipline using the software Protégé4.1whichis a kind of tools to develop ontology. We design there relations between twoconcepts:is_part_of, is_equal_to and is_referenced_by representing hierarchical relation,equal relation and reference relation. We also design the storage method on relationaldatabase so that the ontology database has a good access efficiency.(3)The user interest model is combine with hand-made and model-automated.Firstly the users want to submit their interest manually. The system will update the usermodels automatically by collecting the users’ behavior and content when they browse,download and print the resource. The constructing and updating algorithms based oncontext information are also proposed and designed concretely.(4)We put forward the similarity algorithm between concepts considering therelation between concepts, content weight and the distance between concepts. Thesimilarity algorithm between user interest model and indexing concepts of resource is proposed and accomplish the function of business layer.