Classification of asphyxia & ventricular fibrillation induced cardiac arrest for cardiopulmonary resuscitation

Abstract

In this study we address an important pediatric cardiopulmonary resuscitation problem to identify the cause of a cardiac arrest during the beginning of cardiopulmonary resuscitation. A support vector algorithm was trained and tested using a feature set constructed through wavelet transform analysis of experimental electrocardiography and heart rate data provided by Children's Hospital of Philadelphia. The approach developed in this study yielded an average classification accuracy above 93%.