Modelling and predicting student's academic performance using classification data mining techniques
Raza Hasan, Sellappan Palaniappan, Salman Mahmood, Kamal Uddin Sarker, Ali Hashim Abbas
Abstract
This study focuses on student success analysis and prediction modelling improving the quality in terms of teaching, delivery and satisfaction. Data mining techniques are used to analyse and predict the performance of the student in the specific module or within the timeline of the studies. Supervised learning approach has been adopted with different classification models were tested against the dataset distributed over different levels of study and specialisations. Student grades and online activity on the learning management system were considered as the factors to construct the classifying model. In this study, different algorithms were tested for efficiency and accuracy with the provided dataset for better prediction using WEKA. Random forest was found better and accurate in predicting the student's academic performance. Employing these techniques, it will lead to student preservation and strive for better student satisfaction.