Obstructive sleep apnea (OSAS), which is one of the leading sleep disorders and can result in death if not diagnosed and treated early, is most often confused with snoring. OSAS disease, the prevalence of which varies between 0.9% and 1.9% in Turkey, is a serious health problem that occurs as a result of complete or partial obstruction of the respiratory tract during sleep, resulting in sleep disruption, poor quality sleep, paralysis and even death in sleep. Polysomnography signal recordings (PSG) obtained from sleep laboratories are used for the diagnosis of OSAS, which is related to factors such as the individual's age, gender, neck diameter, smoking-alcohol consumption, and the occurrence of other sleep disorders. Polysomnography is used in the diagnosis and treatment of sleep disorders such as snoring, sleep apnea, parasomnia (abnormal behaviors during sleep), narcolepsy (sleep attacks that develop during the day) and restless legs syndrome. It allows recording various parameters such as brain waves, eye movements, heart and chest activity measurement, respiratory activities, and the amount of oxygen in the blood with the help of electrodes placed in different parts of the patient's body during night sleep. In this article, the performance of PSG signal data for the diagnosis of sleep apnea was examined on the basis of both signal parameters and the method used. First, feature extraction was made from PSG signals, then the feature vector was classified with artificial neural networks, Support Vector Machine (SVM), K-Nearest Neighbors (k-NN) and Logistic Regression (LR).