A Machine Learning Approach for Automatic Student Model Discovery

Student modeling is one of the key factors that affects automated tutoring systems in making instructional decisions. A student model is a model to predict the probability of a student making errors on given problems. A good student model that matches with student behavior patterns often provides useful information on learning task difficulty and transfer of learning between related problems, and thus often yields better instruction. Manual construction of such models usually requires substantial human effort, and may still miss distinctions in content and learning that have important instructional implications. In this paper, we propose an approach that automatically discovers student models using a state-of-art machine learning agent, SimStudent. We show that the discovered model is of higher quality than human-generated models, and demonstrate how the discovered model can be used to improve a tutoring system’s instruction strategy.