Analysis of the Predictive Capacity of Techniques for Modeling Knowledge Applied to Learning Algorithms

• Antonio Carlos Raposo UFMA
• Djefferson Smith Santos Maranhão UFMA
• Carlos de Salles Soares Neto Universidade Federal do Maranhão

DOI: https://doi.org/10.5753/cbie.sbie.2020.1653

Abstract

Learning algorithms is a complex topic and there are several didactic initiatives to improve the student experience, such as intelligent tutors. The predictive capacity of knowledge models is fundamental for the proper functioning of intelligent tutors. Additionally, there are few studies that indicate the effects that the domain has on the predictive capacity of models. This article presents a qualitative and comparative analysis between knowledge models such as Bayesian Knowledge Tracing and Performance Factor Analysis in order to identify possible peculiarities in the field of algorithms. The results show that the compilation errors directly influence the predictive capacity of the models and there are arguments to maintain and to remove them.

Keywords: Predictive Capacity, Knowledge Modeling, Teaching Algorithms

References

Beck, J. E. and Chang, K.-m. (2007). Identifiability: A fundamental problem of studentmodeling. InInternational Conference on User Modeling, pages 137–146. Springer.


Cen, H., Koedinger, K., and Junker, B. (2006). Learning factors analysis–a generalmethod for cognitive model evaluation and improvement. InInternational Conferenceon Intelligent Tutoring Systems, pages 164–175. Springer.


Corbett, A. T. and Anderson, J. R. (1994). Knowledge tracing: Modeling the acquisitionof procedural knowledge.User modeling and user-adapted interaction, 4(4):253–278.


d Baker, R. S., Corbett, A. T., and Aleven, V. (2008). More accurate student modelingthrough contextual estimation of slip and guess probabilities in bayesian knowledgetracing. InInternational conference on intelligent tutoring systems, pages 406–415.Springer.


David, Y. B., Segal, A., and Gal, Y. K. (2016). Sequencing educational content in clas-srooms using bayesian knowledge tracing. InProceedings of the sixth internationalconference on Learning Analytics & Knowledge, pages 354–363. ACM.


Gong, Y., Beck, J. E., and Heffernan, N. T. (2011). How to construct more accuratestudent models: Comparing and optimizing knowledge tracing and performance factoranalysis.International Journal of Artificial Intelligence in Education, 21(1-2):27–46.


Kurup, L. D., Joshi, A., and Shekhokar, N. (2016). A review on student modeling ap-proaches in its. In2016 3rd International Conference on Computing for SustainableGlobal Development (INDIACom), pages 2513–2517.


K ̈aser, T., Klingler, S., Schwing, A. G., and Gross, M. (2017).Dynamic baye-sian networks for student modeling.IEEE Transactions on Learning Technologies,10(4):450–462.


Pavlik Jr, P. I., Cen, H., and Koedinger, K. R. (2009). Performance factors analysis–a newalternative to knowledge tracing.Online Submission.


Psotka, J., Massey, L. D., and Mutter, S. A. (1988).Intelligent tutoring systems: Lessonslearned. Psychology Press.


Ramesh, V. M., Rao, N. J., and Ramanathan, C. (2015). Implementation of an intelligenttutoring system using moodle. In2015 IEEE Frontiers in Education Conference (FIE),pages 1–9.


Raposo, A. C., Maranhão, D., and Neto, C. S. (2019). Analise do modelo bkt na avaliação da curva de aprendizagem de alunos de algoritmos. InBrazilian Symposiumon Computers in Education (Simpósio Brasileiro de Informática na Educação-SBIE),volume 30, page 479.


Yudelson, M. V. (2016). Individualizing bayesian knowledge tracing. are skill parame-ters more important than student parameters?.International Educational Data MiningSociety.


Yudelson, M. V., Koedinger, K. R., and Gordon, G. J. (2013). Individualized bayesianknowledge tracing models. InInternational conference on artificial intelligence ineducation, pages 171–180. Springer