Sleep disorder detection system using support vector machine

Sleep disorders severely impact overall well-being, yet conventional clinical diagnoses remain expensive and uncomfortable, while existing machine learning systems frequently rely on hard to obtain physiological data. To address these challenges, a sleep disorder detection system was developed using the Support Vector Machine algorithm applied to accessible health and lifestyle metrics. Data preparation involved cleaning, normalization, and Recursive Feature Elimination, with testing conducted across several data splits using hyperparameter tuning. The model demonstrated its highest performance with a sixty to forty split, attaining over ninety five percent across accuracy, precision, recall, and F one score when distinguishing between subjects with and without sleep disorders. Ultimately, this system offers an effective early screening tool using simple inputs, providing an accessible preliminary evaluation without replacing formal medical diagnosis.