Augmenting the novice-expert overlay model in an intelligent tutoring system: Using confidence-weighted linear classifiers

In BioWorld, a medical intelligent tutoring system, novice physicians are tasked with solving virtual patient cases. Whilst the importance of modeling and predicting clinical reasoning is recognized, an important aspect of the learner contribution remains unexplored — the written case summary prepared by the learner. The premise of investigating the case summaries is that it captures the thought and process of the learners in solving the cases; since, the case summaries hold important reasoning information, it makes sense to incorporate it as part of the novice-expert overlay model. In this paper, case summaries written by novices and experts were considered as an addendum to the existing novice-expert overlay model in the BioWorld system. Toward this goal, using a promising new classification method called confidence-weighted linear classifiers, this paper proposes a way to augment the novice-expert overlay model in BioWorld.