Research On Ontology-driven Web-based Intelligent Tutoring System

With the fast development of network technology, Internet, as a powerful media, an ultimate method and an essential assistant, is becoming more and more indispensable in area of intelligent tutoring system. To adapt the traditional intelligent tutoring system with E-learning system, and to satisfy the needs of the development towards Web based ITS, are definitely the research trend and also our common mission. Ontology, as one method of knowledge representation, is playing a crucial role in the process of Web based ITS development. In this dissertation, the author proposed a new model of ontology-driven web-based ITS authoring system and implemented the model. By using this method, we can fulfill the needs for meta-model role setup of authoring system, and furthermore, promote the evolution of ITS under the Semantic Web environment. The proper and sufficient usage of ontology is the core of this paper. Also, in order to achieve the target of facilitating the ITS authoring process, we discussed several hot topics including the use of ontology to formalize learner information, domain topics, learning object metadata and pedagogical knowledge.This dissertation first retrospect the research history of ITS, analyzed the different modules and functions within the ITS, and also summarized the current research status and notions of ITS authoring systems. Based on current research status and development technology in the area of ITS, we further discussed the intelligent tutoring system research trend and analysed the potential challenges under the Web and Semantic Web environment, and proposed a creative method of ontology driven authoring for the web based ITS.Next, based on current wide accepted learning technology standards, the author designed the learner ontology, topic ontology, learning object metadata ontology, and instruction ontology used in developing functional modules of the ITS . The learner ontology is based on standards such as PAPI and IMS LIP, and could be used to describe and define the learner character and learning status. Domain topic ontology and learning object metadata ontology are used in domain module setup. The domain topic ontology includes information about topics and descriptions of the relationship between different topics, so that people can compose domain topic map more effectively and sort the topics by their properties or relationships efficiently. Learning object metadata ontology is based on Dublin core metadata, standards of ARIADNE and LOM. It described the properties of different learning objects which serve important factors in resource selection of the instruction process. By connecting these domain topics and learning objects, we then setup the domain module within the ITS.For the instructional module of the ITS, based on the IMS-LD, the author brought out a four-layer instructional design model, including objective layer, event layer, activity layer and resource layer. This model connects the theory and practice, converts knowledge to functional activity structure, provides a solid foundation for ITS instructional strategy design, and also helps the modules authoring of the ITS. Based on the four-layer instructional design model, the author defined the instructional ontology which involved instructional objective ontology, instructional event otology, instructional activity ontology and instructional plan ontology accordingly. Instructional objective ontology is the explicit description of instructional object; instructional event ontology is the description of teaching steps based on certain instructional design theory; instructional activity ontology is the description of detail teaching activities; and instructional planning ontology is the formal expression of every part of the instruction procedure. By utilizing this model, the ITS can customize the instruction strategy with these two kinds of rules: the macro rules for conversion from instructional objective to instructional events, and the micro rules for conversion from instructional event to instructional activities. Further in this dissertation, SWRL based methodology was deployed to express instructional strategy, which could significantly improve the knowledge accumulation, reuse and sharing process.At the end, the author brought in several common ontology representation languages and development tools, and designed a prototype of the ontology-driven web based ITS authoring system. With protege, the author also developed fully functional ontology and instructional rules in SWRL.