Sleep is essential for maintaining mental clarity, emotional stability, and physical health, as supported by extensive research in sleep science. However, modern lifestyles characterized by excessive screen exposure, irregular routines, and heightened stress levels have led to a noticeable decline in sleep quality. Existing AI and digital health tools for sleep monitoring remain expensive or heavily reliant on wearable devices, creating a gap in accessible and non-intrusive sleep assessment methods. To address this issue, the present study proposes a low-cost, software-based AI system that predicts sleep quality without specialized hardware. The system integrates behavioral factors such as exercise duration, caffeine intake, and stress levels with digital wellbeing metrics including screen time, app usage patterns, and nighttime device activity, all of which have been shown to influence sleep patterns. After preprocessing, machine learning models such as Random Forest and XGBoost classify sleep quality into Good, Average, or Poor, aligning with prior research utilizing behavioral and physiological indicators for sleep prediction. A user-friendly dashboard visualizes trends and provides personalized recommendations, such as reducing nighttime screen exposure to improve sleep hygiene. This AI-driven approach offers an accessible and actionable framework for improving sleep health.
Mahajan et al. (Fri,) studied this question.