Comparison Algorithm on Machine Learning for Student Mental Health Data

Abstract

The COVID-19 pandemic has posed unparalleled difficulties, encompassing substantial repercussions on the emotional well-being of students. This study utilises machine learning methodologies to forecast the mental health condition of students during and following the pandemic. The dataset consists of 11 distinct attributes and a total of 101 data points, which have been gathered from multiple sources. The preprocessing stage encompasses the removal of unnecessary characteristics, handling missing data, and partitioning the dataset into separate subsets for training and validation purposes. This study utilises three machine learning algorithms, namely RF, KNN, and NB, in order to make predictions regarding the potential need for psychiatric support among students. These algorithms are carefully optimised to enhance their predictive capabilities. Evaluation metrics commonly used in several fields of study. The findings suggest that the KNN and RF algorithms had outstanding performance, but the Naïve Bayes algorithm exhibited satisfactory accuracy and a balanced trade-off between precision and recall. The optimised models have practical consequences that may be applied at educational institutions and inform policymakers. These implications include the ability to provide tailored interventions and support services specifically designed for students who are facing mental health difficulties as a result of the epidemic. Future research endeavours encompass the need for additional improvement of existing models and the fostering of interdisciplinary collaboration. This study provides significant contributions to the field by examining the utilisation of machine learning techniques in addressing the mental health needs of students both during and after the epidemic.