On Application of Rough Data Mining Methods to Automatic Construction of Student Models

Abstract

Student modeling has been an active research area in the field of intelligent tutoring systems. In this paper, we propose a rough data mining approach to the student modeling problems. The problem is modeled as a knowledge discovery process in which a student’s domain knowledge (classification rules) was discovered and rebuilt using rough set data mining techniques. We design two knowledge extraction modules based on the lower approximation set and upper approximation set of the rough set theory, respectively. To verify the effectiveness of the knowledge extraction modules, two similarity metrics are presented. A set of experiments is conducted to evaluate the capability of the knowledge extraction modules. At last, based on the experimental results some suggestions about a future knowledge extraction module are outlined.