Interrogating Predictive Models to Augment Student Mental Well-Being Through Machine Learning: An In-Depth Exploratory Expedition

Abstract

In order to increase the success rates of therapy, early diagnosis is crucial in mental health care. Every region of the world is affected by the melancholy. This paper's major objective is to use machine learning to psychiatry. By showcasing mental health studies and stressing the limitations of present technology, this study aims to show the promise of ML for health. Due to the potential for profoundly life-altering, long-term effects, the fundamental causes of children's mental health difficulties must be addressed. The processing of medical data now faces challenges that were previously unsolvable thanks to recent advancements in machine learning. When feature selection approaches were used uniformly, the dataset's quality dropped. We examine many methods and put into practice the one that performs best in machine learning prediction models. There needs to be more integration between theoretical and practical machine learning applications in the field of mental health. It is necessary to conduct a thorough inquiry to prevent dismissing irrelevant findings out of hand. The effectiveness of numerous models, including an artificial neural network, logistic regression, AdaBoost, random forest, and a gradient boosting classifier, was examined using a number of metrics. Although the gradient boosting classifier was remarkable with accuracy of 90%, precision of 91%, recall of 85.5%, and an error rate of 0.11, it surpassed them all. The GBC model accurately recognized and predicted the suffering the kids would go through. Academic success, family financial security, and the incidence of violence and bullying on campuses are all powerful predictors of a person's propensity to experience depression or anxiety.