Student model for personalized instructional system
Shihua Huang, Wang Dongqing · 2011
In light of the present situation research on Student model in Personalized Instructional System at home and abroad. This paper discusses a student model using Bayesian network inference as well as cognitive theories to evaluate student's cognition ability precisely and objectively. We set up Bayesian network for expressing a structure of special domain knowledge,and then put the result of student's congnitive ability into Bayesian network ,then use polytree algorithm to update Bayesian network. This approach has been proved to win a more precise characterization since it overcomes the defect, saying independent or discreted, of traditional way which deals with students' information separately. We model the student behavior data, such as scores, in the framework of Bloom's Taxonomy and transmit these evidence to Bayesian network, and in that situation we calculate the cognitive map of student's learning. A course summarization of the algorithm is also discussed. Experimental results show that this approach is effective and practical as in the application.