E-learning attributes selection through rough set theory and data mining

Abstract

Recent research makes wide efforts on attribute selection methods for making effective data preprocessing. The field of attribute selection spreads out both vertical and horizontal, due to increasing demands for dimensionality reduction. The search space is reduced very much by pruning the insignificant attributes. The degree of satisfaction on the selected list of attributes will only be increased through verification of more than one formal channel. In this paper, we look for two completely independent areas like Rough Set theory and Data Mining/Machine Learning Concepts, since both of them have distinct ways of determining the selection of attributes. The primary objective of this work is not only to establish the differences of these two distinct approaches, but also to apply and appreciate the results in e-learning domain to study the student engagement through their activities and the success rate. Hence our framework is based students’ log file on the portal page for e-learning courses and results are compared with two different tools WEKA and ROSE for the purpose of elimination of irrelevant attributes and tabulation of final accuracies.