CSC 591 (Educational Technology) project for credits.

As a part of this project, we are presenting two prediction models to assess student performance for a given set of tasks while learning through an intelligent tutoring system. Support Vector Machines (SVM) are very popular and powerful prediction learning technique so we have made use of them to learn from the knowledge components and step duration for the correct attempt. Bayesian Knowledge Tracing (BKT) is a very popular algorithm used in intelligent tutoring systems to model each learner’s skill mastery. Hence we are running both the models on the KDDCup 2010 Dataset to analyse prediction results by SVM against BKT prediction.

Department of Computer Science, North Carolina State University, Raleigh, NC.