Student classification for academic performance prediction using neuro fuzzy in a conventional classroom

Conventional classroom is still the main learning method applied in undergraduate program of Electrical Engineering and Information Technology Department, Gadjah Mada University. There are several problems in this method, such as large amount of students and limited number of meetings making difficult to understand each student. Student classification is a way to solve the problem by mapping the condition of each student based on certain parameters. Many methods have been applied to classify students that are based on IF-THEN rules and pattern recognition. However, many studies were done on intelligent tutoring systems and e-learning systems, not in a conventional classroom. Moreover, there are no researches that measure basic values by considering intelligence and non-intelligence performances. In this work, a student classification model was developed by applying neuro fuzzy concept; a combination of fuzzy's IF-THEN rules and neural network's ability to learn, so this method has the ability to learn from the generated rules to produce the best classification model. The model can be used to predict students' academic performance. Data were processed using ANFIS Editor-Matlab Fuzzy Logic. The results showed that combination of three parameter values -interest, talent, and motivation- is the best model for students classification, which has training RMSE value 0.12301 and testing average RMSE value 0.25611.