A Prediction System for the Effect of Electrical Defibrillation Based on Efficient Combinations for Feature Parameters

In this paper, we propose a prediction system of the effect of electrical defibrillation for shockable arrhythmias. In order to develop the proposed prediction system, ECGs are firstly analyzed by Gabor wavelet transform, Poincare plot and spectral entropy, and feature parameters are extracted by these analysis results. Moreover, the imbalanced data are corrected by using SMOTE (Synthetic Minority Over-sampling TEchnique), and we adopt Pearson’s χ2 test so as to evaluate the efficient feature parameters. Finally, by using selected feature parameters, support vector machines (SVM) based on three kernels (Linear, Gaussian, and Polynomial) are constructed, and the effectiveness of the prediction system is presented.