On the Efficiency of Multiple Linear Regression over Artificial Neural Network Models

There has been a considerable and continuous interest to develop models for rapid and accurate modeling of students’ academic performances. In this study, an Artificial Neural Network model (ANNm) and a Multiple Linear Regression model (MLRm) were used to model the academic performance of university students. The accuracy of the models was judged by model evaluation criteria like and The modeling ability of the developed ANN model architecture was compared with a MLR model using the same training data sets. The squared regression coefficients of prediction for MLR and ANN models were 0.746 and 0.893, respectively. The results revealed that ANN model proved more accurate in modeling the data set, as compared with MLR model. This was because ANN model had its as against the traditional model which it’s was 0.182. Based on the results of this study, ANN model could be used as a promising approach for rapid modeling and prediction in the academic fields.