Machine Learning Approach to Analyze Factors Affecting Student Mental Health

The growing epidemic of mental health problems among students has been acknowledged as a global health crisis, particularly in educational settings. Thorough investigation is essential to develop effective interventions that improve students’ overall mental well-being. This study aims to analyze a dataset that contains various factors that influence mental health, identify the most disruptive attributes, and uncover those that are generally overlooked or considered insignificant. We have applied data-augmentation techniques to achieve the research objectives of this paper. This data set required data augmentation to acquire the expected significant attributes. A unique dual correlation analysis approach was employed, using Chi-Square and Pearson Correlation tests, to investigate the relationships between categorical and continuous variables. Limitations in data representation led to dataset refinement, focusing on the dominant 18–26 age group, followed by a second analysis that revealed clearer insights into key contributing factors. Additionally, Random Forest and Naive Bayes machine learning models were applied, with Random Forest demonstrating superior performance, achieving an accuracy of 75.85%, compared to 71.68% for Naive Bayes.