A Research Of Error-correction Diagnosis Model Based On Agents

Intelligent Agents is featured with autonomy, sociability, reactivity and pre-activeness, which has given rise to widely applying of agents in the fields of Intelligent Tutoring System. According to the Constructivism learning strategies, the distance education system based on Web should not only design learning environment properly, but also attach importance to the learning process of the learners, especially should think much of errors and mistakes that learners make in their learning process, and reasons and influences. Therefore, the author took on the application of agents in ITS as the entry, then studied theories on error-correction diagnosis model in ITS, and proposed the constructing methods for diagnosis model archetype system on the basis of agents. Combined with the archetype, the author also presented her diagnosis algorithm based on intersection. In order to verify this error-correction diagnosis model, the author developed HSK Intelligent Networked Tutoring School based on research results about single agent application and diagnosis algorithm. The HSK ITS has been given an appraisal advanced level in the world of by Beijing Technology Committee. In addition, the author did an one-factor-two-level teaching experiment of small size, which proved that HSK ITS could diagnose and correct the learners' errors, build up their adaptability, and show remarkable effect on learners. So the author made the further study on multi-agents applying to distance education, and also probability applying to diagnosis algorithm. A multi-agents archetype system of diagnosis model was proposed. The subsequently running test indicated that the structure of diagnosis MAS is more reasonable, and that the improved diagnosis algorithm is more effective.In Section One, the author described a series of concepts and theories about teaching diagnosis, and addressed the major function and important effects that diagnosis model played in distance education.Section Two presented agents' meaning, characteristics and related theory systemically, and many theories and technique that can be applied to diagnosis model have been described. Here the author also discussed their advantages and disadvantages when using for diagnosis. Then the author describes the conclusions made in previous work, and analyzed the relation between agents and ITS, and considered that implementing diagnosis model based on agents is feasible and available.In Section Three, the author proposed a error-correction diagnosis model based on agents, and described the framework and diagnosis algorithm. The paper also addressed the developing process and some key skills involved in archetype in detail.Finally, the author pointed out the improvements that her study needed according to her research achievements. How to make the diagnosis model and agents technology play effective role in networked ITS, which need further study and experiments.