Discovering hierarchical patterns of students' learning behavior in intelligent tutoring systems

Abstract

We present a granular approach to discover new and interesting learning behavior patterns of students learning with an intelligent tutoring system. An n-gram analysis is used to model the learning behavior from learning action streams. The regular and irregular learning behavior patterns are obtained from the n-gram analysis model. Then, the n-gram models are clustered into a hierarchy. The hierarchical pattern can be used to improve the domain knowledge of an ITS in predicting student's actions, sequencing problems to be solved, and adjusting hint mechanisms. Our approach is domain independent and able to manage learning behavior uncertainties. © 2007 IEEE.