Feature Reinforcement in Intelligent Automated External Defibrillators for Sudden Cardiac Arrest Detection

Sudden cardiac arrests are caused by shockable rhythms known as ventricular fibrillation and ventricular tachycardia. Rapid diagnosis implemented by the automated external defibrillation results in electrical shock, which improves the chance of survivals. In this paper, a novel method is developed to design an effective shock advice algorithm in the automated external defibrillation. An optimal set of 15 features are selected carefully by the feature selection algorithm using K-nearest neighbors and the fuzzy C-mean clustering, which produces reinforced features. Various machine learning methods are considered for the performance estimation of the optimal feature set and entire input features using cross validation procedure. The simulation results, which are accuracy of 99.01%, sensitivity of 99.14%, specificity of 98.97, show that the proposed shock advice algorithm for the automated external defibrillation is potential for practical application in real clinic environment.