Integrating Deep Learning into Educational Wellbeing: Early Screening of Anxiety, Depression, and Stress Among University Students

The growing prevalence of anxiety, depression, and stress among students highlights the urgent need for school-based strategies that promote psychological well-being and timely intervention. This study explores the use of artificial intelligence (AI) as a scalable and data-driven tool to support institutional mental health initiatives in higher education. Using synthetic and real datasets derived from validated psychometric instruments (the Beck Anxiety Inventory (BAI), Beck Depression Inventory (BDI), and the Perceived Stress Scale (PSS-14)), we trained and evaluated 32 deep neural network architectures for the early detection of emotional distress. Optimized three- and four-layer dense models achieved classification accuracies exceeding 95%, demonstrating the feasibility of deploying AI-based screening tools in educational settings. Beyond prediction, this approach can support counselors and educators in identifying at-risk students and informing proactive, school-based interventions to improve mental health and resilience in post-pandemic academic environments.