Examination scores: prediction using artificial neural network (ANN) / Nabilah Ismail

This paper presents a model for predicting the final examination scores using an artificial neural network. The scope of this paper is to distinguish the components affecting the performance of students in final examinations and to predict the grade in the final examination. A sample of 112 students from the Faculty of Electrical Engineering, UiTM Shah Alam who have attempted at least once for the Power Engineering course was tested and trained. The students’ assessments were used as the inputs to train the ANN model. The code was written and executed using MATLAB format. A method of Levenberg-Marquardt and Gradient Descent were used as a training algorithm and the performances were compared in term of accuracy. The results showed that the model is able to correctly predict the examination score with an accuracy of 94.06%.