Machine Learning for Detection of Obstructive Sleep Apnoea

Abstract

Obstructive sleep syndrome (OSAS) has become an important health concern. It can have adverse consequences on brain function, mood, and productivity at work that hampers the quality of life and increased risk of occupational injuries. Polysomnography (PSG) is considered as an established and effective diagnostic tool for providing information on the severity of OSAS and the degree of sleep fragmentation. However, the numerous steps in the test to diagnose obstructive sleep syndrome are expensive and time-consuming. This study is aimed to visualize the potency and clinical relevance of various machine learning algorithms supported on demographic data and form knowledge to predict obstructive sleep syndrome severity.