Neural Network Prediction Of SPM Achievement

The purpose of this study is to build a neural network model for prediction of SPM achievement for the students in a Malaysian secondary school. The neural network model uses multi-layer perceptron involving a backpropagation algorithm and the tangent sigmoid as the transfer function. This study does not only consider the students’ grades for the core subjects that they take in the SPM but
also the student gender. Based on the model results, the real exam performance is to be predicted. This study shows that neural network can be trained with students’ data to predict their achievement in the SPM examination.