A GRAPH BASED APPROACH TO LEARNER PROFILING IN AN INTELLIGENT TUTORING SYSTEM

Abstract

Learner profiling lays the foundations of the personalization that happens in adaptive educational applications and Intelligent Tutoring Systems (ITS). A learner’s level of knowledge on a topic is estimated from their performance on certain activities related to the topic. For this, researchers have devised many model extensions throughout the years that incorporate specific cognitive features into student profiling. In this paper, a new graph-based algorithm for learner profiling has been proposed that is able to adapt the course to the current knowledge level of the learner using the topic dependencies fed in by the subject experts and the past response data of learners who have taken this course in the past. This results in learner profiling with minimum number of assessment activities in the best case.