The Research On The Intelligent Tutoring System Based On The Web And Data-mining

With the development and application of technology, more and more educators put emphasis on the Intelligent Tutoring System. In order to improve the extent of intelligence the application of artificial intelligence is mostly given prominence to. This research is expected to discuss the operational principle of intelligent system, analyze the key steps in constructing the system and the deficiency in the traditional system, and discuss the principle of the technology of Data-mining and its application in the intelligent system. Based on that, the author put forward a new structure, basic strategy and method for constructing the student-model-centered system. The main content of this thesis is divided into four parts: Part 1: Preface.Through reviewing the development of ITS at home and overseas, we realize the level of its development and find out that the majority of web-based instructional systems are presenting the problem of low-level intelligence, although they are bestowed to the design idea and developmental pattern of ITS. In the web-based instructional systems which supported by database technology based on student information base, knowledge base and the base for instructional strategy, there is comparative large space in the potential information mining on the web record of students' learning behaviors in the web-based activity. So the author bring forward this research. Part 2: The research on Intelligent Tutoring System (ITS).In this part some basic concepts are defined, and indicate the traditional ITS is consisted of four components: student model, instructor model, knowledge base and user interface model. Student model is the foundation of the whole system. Instructor model executes the instruction. In this type of system, instructor model performs the reasoning function depending on knowledge base and student model. The key techniques for ITS include the selection of knowledge representation, reasoning mechanism, the construct of the instructor model and the reasoning based on collaborative components. The methods of knowledge representation include the frame knowledge representation, the semantic network knowledge representation and the production rule knowledge representation, which have their own merits and limitations.