Title: Determining fuzzy rules for student's performance and learning efficiency by using a hybrid approach
Authors: Norazah Yusof, Abdul Razak Hamdan
Addresses: Fakulti Sains Komputer & Sistem Maklumat, Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia. ' Fakulti Teknologi & Sains Maklumat, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor, Malaysia
Abstract: This paper describes a hybrid approach that combines a fuzzy inference system with a neural network, and with a rough set technique in determining the fuzzy rules from a fuzzy rule base system of the student model. The back-propagation neural-fuzzy approach is used to solve the problem of incompleteness in the decision made by the human experts. By training the neural network with selected patterns that are certain, the proposed approach was expected to produce decisions that could not previously be determined, and accordingly, a complete fuzzy rule base is formed. This paper proposes a rough-fuzzy approach that reduces the complete fuzzy rule base into a concise fuzzy base. After comparing the defuzzified values of the complete fuzzy rule base with the concise fuzzy rule base, it is discovered that the performance of the concise fuzzy rule base does not degrade and it remains complete and consistent.
Keywords: student models; hybrid approach; neural networks; rough-fuzzy; fuzzy inference systems; fuzzy rule base; rough sets; intelligent tutoring systems; student performance; learning efficiency; student learning.
International Journal of Reasoning-based Intelligent Systems, 2010 Vol.2 No.2, pp.142 - 157
Available online: 30 Aug 2010 *Full-text access for editors Access for subscribers Purchase this article Comment on this article