Probabilistic Classifiers and Statistical Dependency: The Case for Grade Prediction
The research presented here aims at predicting grades by means of a set of relevant student’s variables. To solve this problem, a probabilistic approach was applied in which we assume that the probability of obtaining a certain grade is conditioned on the personal attributes of each student. A Bayesian classifier was the natural choice to include the student’s attributes in the estimation of the likelihood. However, a striking result was observed when the accuracy of the bayesian prediction was lower than the one provided by a baseline predictor based on student’s clustering. A follow-up analysis explains the reason behind this result and provides a guideline for similar classification problems.
KeywordsProbabilistic models Bayesian classifier Statistical dependency Education
This work has received financial support from the Ministry of Science and Innovation of Spain under grant TIN2014-56633-C3-1-R as well as from the Consellería de Cultura, Educación e Ordenación Universitaria (accreditation 2016–2019, ED431G/08) and the European Regional Development Fund (ERDF).
- 2.Edin, O., Mirza, S.: Data mining approach for predicting student performance. Econ. Rev. – J. Econ. Bus. X(1), 3–12 (2012)Google Scholar
- 4.Gedeon, T.D., Turner, H.S.: Explaining student grades predicted by a neural network. In: Proceedings of 1993 International Joint Conference on Neural Networks, IJCNN 1993, Nagoya, vol. 1, pp. 609–612 (1993)Google Scholar
- 6.Kotsiantis, S.B., Pintelas, P.E.: Predicting students marks in Hellenic Open University. In: Fifth IEEE International Conference on Advanced Learning Technologies, ICALT 2005, pp. 664–668 (2005)Google Scholar
- 7.Qasem, A.A., Emad, M.S., Mustafa, I.N.: Mining student data using decision trees. In: International Arab Conference on Information Technology (2006)Google Scholar
- 9.Superby, J.F., Vandamme, J-P., Meskens, N.: Determination of factors influencing the achievement of the first-year university students using data mining methods. In: Proceedings of the Workshop on Educational Data Mining at the 8th International Conference on Intelligent Tutoring Systems, pp. 37–44 (2006)Google Scholar