Bayesian diagnosis tracing model (BDT) replaces the generic “wrong” response in the classical Bayesian knowledge tracing model (BKT) with a vector of procedure misconceptions. Using a novel dataset with actual student responses, this paper shows the BDT model has better interpretability of the latent factor and minor improvement in out-sample predictability in some specification than the BKT model.
CITATION STYLE
Feng, J., Zhang, B., Li, Y., & Xu, Q. (2019). Bayesian diagnosis tracing: Application of procedural misconceptions in knowledge tracing. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11626 LNAI, pp. 84–88). Springer Verlag. https://doi.org/10.1007/978-3-030-23207-8_16
Mendeley helps you to discover research relevant for your work.