<![CDATA[Sleep disorder is a disturbed breathing flow (collapse) during sleep. The symptoms are generally undiagnosed and untreated properly so that repeated respiratory interruptions have the potential for severe sleep disorders. Electrocardiogram (ECG) recordings are practical tools used to examine the existence of sleep disorders in the heart rhythm. The ECG represents heart electrical activity in the form of P, QRS, and T waves. The number of ECG sensors is uncomfortable for the patient to record the data, increasing the recording complexity, slowing the computation, causing misinterpretation and loss of clinical information. Therefore, an early warning system is needed as a medical aid that can be diagnosed using single-lead ECG. In conducting this study, the system consists of five stages, which include the acquisition of ECG records, pre-processing, extraction of features, selection of features, and the classification process. ECG-record feature sets consist of time-domain, frequency-domain, and non-linear analysis. The AdaBoost method confirms that the model had the highest performance than the SVM, k-NN and NN. The results of the experiments thus measure the outperformed of method performance and achieved 90.1% classification accuracy for the AdaBoost classification method. Moreover, the F1 score, precision, recall, sensitivity, and specificity was reported as 90.1%, 90.3%, 90.1%, 86.9%, and 93.3%, respectively.]]>
Copyrights © 2020
