Machine Learning Prediction of Student Academic Performance Based on Educational and Lifestyle Factors

Abstract

This study aims to investigate the impact of lifestyle factors on student academic performance, addressing a critical gap in the existing literature. The research seeks new insights to improve student academic performance, contributing to a deeper understanding of educational prediction based on academic and lifestyle factors. The study employs a machine learning approach, using ML algorithms such as Random Forest, Decision Tree, Bagging, KNN, Naive Bayes. The dataset was downloaded from the UCI Machine Learning Repository and then processed with MS Excel. The results indicate that five machine learning algorithms namely, K-Nearest Neighbor (KNN), Naive Bayes (NB), Bagging, Random Forest (RF) and Decision Tree (DT) were used. The value of the systemic analysis and its use of different performance variables allowed the manuscript to show how different academic and lifestyle variables combined leading to the predictive ability of the algorithms. Our model with Random Forest algorithm produced 99.23% accuracy, which is better than the 91.37% accuracy when only academic features were used, as reported in the baseline study. The present findings support the hypothesis that both academic and lifestyle variables have significant influence on the improvement of academic performance predictive models. By the implementation of machine learning to predict student academic performance, this study produces a new point of view to predict academic performance based on education and lifestyle variables and contributes to the existing knowledge of educational outcomes. The results have implications for implementing intelligent educational analytics and suggest topics for further research, policy and practice implications.