Multimodal ML Framework for Predictive Chronic Stress Assessment in Student Populations

computer science

Abstraction Chronic stress of students has emerged as a major issue as it affects the academic outcome, mental stability, and the overall health of students to a considerable degree. The complexity and dynamism of the way stress develops are usually not captured in the traditional types of diagnosis. This paper proposes the AB Method (Anub), a machine learning architecture that can identify and forecast chronic stress in students according to the binary data of their symptoms, which include sleep problems and irritability, poor diet, etc. The training data was comprised of 80 students whose symptom counts of five or above gave a diagnosis of the chronic stress. The methodology we used combines preprocessing, feature remodeling, symptom duration forecasting, and simulation of the symptom trend in predicting stress over five years ahead. The model is not only quite accurate but also justifiable in terms of the importance analysis of the features and allows personalized interventions. The predictive simulations make stakeholders know how stress can develop in individual students over time. Results indicate that the AB Method is able to provide meaningful comprehension on important contributing symptoms and generate interpretable longitudinal predictions. The model delivered good performance with a clear visualisation and early warning. The implication of this study is that interpretable AI-based models such as AB can provide scalable and proactive solutions to educational mental health programs that move towards preventive intervention of caregivers rather than responsive care.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.