A Machine Learning-based Framework for Sleep Disorder Classification using Decision Tree Algorithm and Streamlit Deployment

Sleep disorders significantly impact global health, affecting millions worldwide through conditions such as insomnia and sleep apnea. This research presents a comprehensive machine learning pipeline for automated sleep disorder classification using the Sleep Health and Lifestyle dataset. Our approach employs a Decision Tree classifier integrated with robust preprocessing techniques including feature scaling, categorical encoding, and missing data handling. The system incorporates stratified sampling to maintain class distribution and implements comprehensive evaluation metrics including confusion matrices, classification reports, and accuracy assessments. The model achieved training accuracy of 89% and testing accuracy of 84%, demonstrating effective generalization capabilities. Key preprocessing components include safe transformation functions for handling unknown categories, Min-Max scaling for numerical features, and systematic feature alignment. The research contributes to healthcare informatics by providing an interpretable, scalable solution for early sleep disorder detection based on lifestyle and demographic parameters. Visualization components including confusion matrix heatmaps and performance comparison charts enhance clinical interpretability. The findings demonstrate the potential of machine learning algorithms in supporting clinical decision-making for sleep disorder diagnosis and management.