A Temporal and Classification Approach to Predicting Stress and Work-Life Balance Among Higher Education Teachers

Elevated stress levels among higher education faculty, caused by increased workload, administrative responsibilities, and academic responsibilities, have severe consequences for both individual well-being and institutional efficacy. If left unaddressed, this stress result in burnout, decreased job satisfaction, and impaired work-life balance, negatively impacting educators' performance and, consequently, the quality of education. Resolving this issue is essential for cultivating supportive and effective educational settings. This study presents a prediction algorithm that integrates temporal and classification analysis to effectively evaluate and forecast stress levels in educators. Through the examination of workload patterns, teaching hours, and self-reported stress levels, the model attains exceptional performance metrics, including 96.7% accuracy, 0.95 F1 score, and 0.94 recall, surpassing previous models. The model design employs dense and LSTM layers to proficiently capture both static and temporal trends, reinforced by batch normalisation and dropout layers to mitigate overfitting. This method provides a notable advancement compared to conventional methods, boosting forecast precision and dependability. The results indicate that this model can offer educational institutions a powerful instrument for tracking stress and facilitating early treatments, therefore aiding teachers' work-life balance and promoting healthier, more productive educational environments.