Discipline-Ontology Based Learning Resources Semantic Retrieval Algorithm

Abstract

There is much difference among learners and learning resource providers when the semantic of learning resources is understood and expressed, and it is the important reason which causes lower accuracy in learning resources retrieval. In order to resolve the problem above, three mechanisms are applied as follows: 1) Discipline Ontology is constructed, which is the formalization for concepts and the relationships between concepts existing in some discipline domain. OWL is adopted as Discipline Ontology description language; 2) Inference rules are defined on the basis of Discipline Ontology. Semantic extension on keyword from user is performed by using Jena inference engine and inference rules , so as to better interpret and describe the requirement; 3) Learning resource metadata is extracted and defined by following Learning Resource Meta-data Specification, so as to provide formal description for learning resources. A semantic retrieval framework for learning resources is presented, and the process of learning resource semantic retrieval algorithm is discussed in detail. Firstly, the semantic extension on inquiry keyword from user is performed on the basis of Discipline Ontology; secondly, by using the improved similarity calculating formula, the keywords produced by semantic extension are sequenced. A set of keywords which have higher similarity with inquiry keyword are sorted out, and are used as inquiry keywords; then, search is performed on the basis of inquiry keywords and learning resource metadata. A set of descriptions for learning resources, which probably meet the requirement of user, is sent to user. The algorithm provides an approach for learning resource retrieval, and is able to support the effective access on learning resources.