Early Detection of College Student Mental Health Issues Using Machine Learning

This paper explores using machine learning to create early warning models that identify college students at risk for mental health issues like depression and anxiety. By analyzing data such as academic performance, demographics, campus involvement, and digital activity, we aim to proactively connect at-risk students with mental health resources. We apply both supervised and unsupervised techniques, including classification, clustering, and neural networks, to a dataset from a large public university. The best-performing model, random forest, achieved 81% accuracy in identifying students with risk factors for depression and anxiety. Clustering analysis revealed common patterns of academic struggles, social isolation, and addictive digital behavior among at-risk groups. These findings highlight the potential for machine learning to support early intervention and improve student well-being and success. Future work will focus on refining the models, expanding to additional campuses, and integrating predictive tools with university mental health strategies.