Lifestyle to Sleep Health: A CNN-LSTM Approach for Predicting Sleep Quality and Disorders

This paper explores the impact of lifestyle factors on sleep patterns, highlighting its public health significance. By integrating online and real-life data, the paper investigates how lifestyle elements such as physical activity, stress, and health conditions influence sleep quality and the prevalence of sleep disorder. Using statistical analysis, the study identifies key predictors of sleep quality. It reveals that Sleep Duration and Stress Level are the strongest factors associated with sleep quality, while elements like Daily Steps and Blood Pressure contribute minimally. This study proposes a predictive model based on a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) architecture, optimized via hyperparameter tuning. This model achieves an initial accuracy of 86% in detecting sleep disorders and a mean squared error (MSE) of 0.77 for sleep quality prediction. These results underscore the importance of modifiable lifestyle behaviours in influencing sleep, suggesting strategies to improve sleep hygiene. The model features a user-friendly interface where users input their lifestyle details. It processes this information to predict sleep quality and disorder status, then provides personalized recommendations for improvement. The study demonstrates how machine learning and statistical analysis can provide insights into sleep health, offering a foundation for future interventions aimed at optimizing sleep quality through lifestyle changes.