Machine Learning-Based Prediction of Student Academic Performance

Abstract

One of the important applications of educational data analytics and machine learning is predicting students' academic performance. Schools produce a great deal of information regarding students – their attendance, grades, past achievement and behaviour, results of assignments and assessment marks, etc. – that can be analysed to uncover patterns linked to academic success. The current study used the predictive analytics approach for creating and testing machine-learning models for the prediction of students' academic performance in universities. A dataset devoid of names and identities that included attendance, previous grades, hours studied, assignment grades, and assessment scores was gathered from an educational institution. The data was cleaned and normalised prior to split of the data into training and testing sets. Decision tree algorithm, random forest algorithm and support vector machine were used to train and evaluate three machine-learning algorithms. Accuracy, precision, recall, F1-score and mean absolute error (MAE) were used for comparing the model performance as appropriate. The purpose of the study was to determine which algorithm was the most accurate in predicting student academic performance. The results would be expected to show the promise of machine learning for early identification of students at academic risk and how it helps education systems to tailor effective, timely interventions.