Pre-determination of OSA degree using morphological features of the ECG signal

摘要: 30 OSA patients were automatically classified using electrocardiogram (ECG) data.In total, 29,127 epochs identified as mild, moderate, and severe.Fifteen morphological features extracted from these epochs.Success rates of 97.202.15% 90.188.11% with the SBFS algorithm obtained.ANN, NB, RF, DT, LOGR SVM classifiers used to obtain best result. Obstructive sleep apnea (OSA) is a very common, but difficult disorder diagnose. Recurrent obstructions form in airway during sleep, such that can threaten breathing capacity patients. Clinically, continuous positive pressure (CPAP) most specific effective treatment for this. In addition, must be separated according its degree, CPAP applied this study, two different databases data, severe. One was original recordings which had 9 8303 other one Physionet benchmark database 21 20,824 epochs. Fifteen could when seen, both before after it presented. Five data groups total first dataset second prepared 10-fold cross validation effectively determine test data. Then, sequential backward feature selection (SBFS) understand more features. The evaluated artificial neural networks (ANN) optimum classification performance. All processes repeated ten times error deviation calculated accuracy. Furthermore, are frequently literature tested selected degree estimated three pre-apnea yielding success datasets, respectively. Also, classifier followed ANN system 96.233.48% 88.758.52% datasets.