Predicting students’ grades using fuzzy non-parametric regression method and ReliefF-based algorithm
Abstract
In this paper we introduce two new approaches to predict the grades that university students will acquire in the final exam of a course and improve the obtained result on some features extracted from logged data in an educational web-based system. First we start with a new approach based on Fuzzy non-parametric regression; next, we introduce a simple algorithm using ReliefF estimated weights. The first prediction technique is yielded by integrating ridge regression learning algorithm in the Lagrangian dual space. In this approach, the distance measure for fuzzy numbers that suggested by Diamond is used and the local linear smoothing technique with the cross validation procedure for selecting the optimal value of the smoothing parameter is fuzzified to fit the presented model. Second approach is based on ReliefF attribute estimation as a weighting vector to find the best adjusted results. Finally, to check the efficiency of the new proposed approaches, the most popular techniques of traditional data mining methods are compared with the presented methods.
Keywords
Educational Data Mining; Predicting Marks; Fuzzy Non-parametric Regression; KDD; ReliefF; WEKA; Matlab