Owing to the increasing demand of learning resource sharing, data integration is a promising approach to support system interoperability across heterogeneous e-Learning systems. One of the main problems to deal with in data heterogeneity is semantic conflict and although a variety of methodologies have been proposed which focus on name conflicts, structure conflicts have received little attention. This paper proposes an approach of ontology-based metadata integration for sharing learning resources between the heterogeneous Learning Management Systems. In order to achieve semantic interoperability, the standard metadata DC/LOM, WordNet and mapping rules are proposed to cope with not only name conflict but also structure conflict arising in the real-world system. The preliminary result provides the ability of learning resource interoperability between the different repositories that can establish a community-shared and reusable semantic pattern base over educational institutes.
I. INTRODUCTIONThe need for the sharing of learning resources is a promising approach to provide data interoperability between the different repositories. It is relevant to a number of heterogeneous e-Learning systems. These heterogeneities are of concern to both architecture and data. Since an e-Learning system provides learning materials for the users who attend a course, it still provides different content formats such as text, image, audio, or video as well as a variety of metadata. Therefore, data integration is a possible approach in overcoming data heterogeneity.Ontology has been extensively used in data integration systems [1]; however, semantic conflict is one of the main sources of problems amongst heterogeneous ontologies. For this reason, semantic reconciliation is a major challenge for integration techniques in order to reach interoperability. There are two problems [2] which occur when the information is integrated over different ontologies. The first is metadata heterogeneity and is concerned with the meaning of described information; the second problem is instance heterogeneity and involves the different representations of instance values in the same ontology (this is also known as representation conflicts). This conflict needs to be addressed before ontology interoperation takes place. This paper concentrates on the first problem i.e. metadata heterogeneity.There are two types of conflicts related to metadata heterogeneity and they are classified as follows: