Sleep Disorders Detection and Classification Using Random Forests Algorithm

Insomnia and Sleep Apnea are popular sleep disorders. Sleep detection is an important step, especially in the earliest diagnosis of mental disease analysis. Moreover, sleep disorders affect body health such as blood pressure and stroke. Traditional detection methods are expensive and time-consuming due to devices required to read signals and experts for understanding and analyzing these signals. Therefore, different automatic systems based on machine learning algorithms have been developed to detect sleep disorders based on pre-assembled data from different clinics. In this chapter, a sleep disorders forecasting model is implemented using a Random Forests Classifier algorithm. The model is trained and tested using Sleep Health and Lifestyle dataset. The Sleep Health and Lifestyle dataset includes three classes Insomnia, Sleep Apnea, and None. Each class is featured with 12 different values such as gender, Sleep Duration, and Quality of Sleep. The detection accuracy of the Random Forests Classifier algorithm is recorded to be 88% on the sleep Health and Lifestyle dataset. Moreover, different algorithms were trained and tested on the same dataset to measure the performance of the selected algorithm. The result showed that the Random Forests Classifier algorithm is better than the other algorithms.