Study And Its Application Of Ontology-based Discovery Mechanism For Learning Objects

With the rapid development of the technologies related with computer network and computer application, the technologies have impacted on many domains. Internet has become a popular tool for people to acquire knowledge and communicate. It is a trend to unite the Information Technology and other kinds of technology including pedagogics. As a result, there have been various of learning objects accessible via the Internet. Sharing these learning objects with different parties and different systems is a good way to enhance the level of teaching and lower the teaching cost. However, there are two barriers. Firstly, semantic searching on learning objects isn't feasible without a unified presentation model. Secondly, the heterogeneity of distributing learning objects makes the seamless integration of learning objects impossible.Semantic Web and Service-Oriented Architecture (SOA) throw light on the integration of heterogeneous resources. The dissertation focuses on importing Semantic Web and SOA into the discovery mechanism of learning objects. It is worthwhile to apply Semantic Web and SOA to pedagogics domain. The dissertation proposes the discovery mechanism of learning objects based on SOA and Semantic Web by combining the ontology and metadata standard so as to figure out the sharing problem about the distributing, heterogeneous learning objects in totally new view. The work and achievements are as follows:(1)The discovery mechanism of learning objects based on SOA and Semantic Web is proposed.The dissertation validates the feasibility of applying the technologies related with SOA to manage the learning objects. The solution extends the range of Services in common sense to the all kinds of resources available on the Internet, such as different kinds of data, information, applications, hardware and so on. By adding internal semantic pattern and importing external semantic pattern, the semantic description for Web Service is achieved in the discovery mechanism; meanwhile an external matching engine is designed to implementation the semantic match in SOA. Therefore, not only is the obstacle in the discovery of distributed learning objects removed, but also a new path is broken for the discovery of different kinds of resources on the Internet.(2)The learning object ontology based on the learning object metadata is built.Combining the standards for the learning object metadata and ontology technology brings us the sharing of the learning objects. At first, the dissertation analyses the Standard IEEE LOM, and then discusses the possibility and feasibility of employing the standard during ontology development. The core elements of the Standard IEEE LOM are used as the properties in the learning object ontology. By this mean the learning object metadata bridges the learning object and the ontology. The solution makes ontology building much easier, at the same time the ontology is more general and extensive. It is signification for applying the ontology technology into the education domain.(3)The incremental methodology for domain ontology is studied.Based on the similarities between the development of software and ontology, the incremental methodology is presented in accordance with the rules of ontology development. More than one discipline of software development is employed in this methodology, which comes from the engineering approach. According to the methodology, the increment process model is used in the ontology development process, which is a classic model in Software Engineering. Object-Orientation (00) characterizes the methodology. The 00 modeling approach can help to organize the hierarchical structures during the ontology development because the notation of class in ontology is similar to that in OO. Ontology for Learning Objects and ontology for Web Services are built with the methodology in the dissertation.(4)The matching algorithm supporting semantics and Quality of Services (QoS) is proposed.The dissertation focuses on how to match the learning object with semantics. In order to achieve the goal, the dissertation proposes a multiple levels matching algorithm in sequence: matching by text, matching by semantics and sorting by QoS. Matching by text is based on word similarity, which differs from most existing algorithm. By this way, the computing complexity is much lower and the matching precision is kept. The dissertation analyses the factors which influence the semantics in ontology and makes full use of them in matching by semantics. The factors are the depth of concept, the relation of different concepts and difference of property between concepts. Sorting by QoS characterizes the matching algorithm. The algorithm is based on fuzzy mathematics and the result is displayed in similarity.