A Comparative Analysis of Machine Learning Models for Predicting Mental Health of At-Risk Students

This research examines the efficacy of machine learning algorithms in forecasting mental health difficulties, specifically depression, anxiety, and panic attacks, within the university students’ population. Utilising a publicly available dataset from Kaggle, the study evaluates the predictive performance of four prominent models: Random Forest, Logistic Regression, Naive Bayes, and XGBoost. Model performance is assessed based on key evaluation metrics, including accuracy, precision, recall, and F1-score. XGBoost emerged as the most effective model, achieving 98% accuracy and high recall, highlighting its potential for identifying at-risk individuals in imbalanced datasets. The study also identifies key predictors, including depression and anxiety indicators, while emphasising the importance of scalable, data-driven interventions in educational settings. The findings contribute to the growing field of AI-based mental health prediction, offering insights for early detection and tailored support systems. Keywords: Data-Driven Interventions, Imbalanced Datasets, Ensemble Learning, Mental Health Prediction, XGBoost. Proceedings Citation Format Fatimah Adamu-Fika, Isaiah Ibukunoluwa Adesokan, Henry Onyeoma Mafua, Aisha Tijjani Ramalan, Ahmed Taiye Mohammed, Oluwaseyi Ezekiel Olorunshola, & Onyinye Vivian Okpoko (2023): A Comparative Analysis of Machine Learning Models for Predicting Mental Health of At-Risk Students. Proceedings of the 37th iSTEAMS Multidisciplinary Cross-Border Conference. 30th October – 1st November, 2023. Academic City University College, Accra, Ghana. Pp 211-226. dx.doi.org