Arrhythmia Classification Using Serial Fusion of Support Vector Machines and Logistic Regression

摘要: Reliable arrhythmia classification from complex electrocardiogram (ECG) signals is one of the most challenging pattern recognition problems. Several individual classifiers have been studied in ECG domain. Also, parallel and serial classifier fusion systems proposed to increase reliability. In this study, we are mainly interested producing high confident results be applicable diagnostic decision support systems. We first experiment compare two common techniques: vector machines (SVM) logistic regression (LR). Then, propose a two- stage system based on SVM's rejection option. relate distance outputs confidence measure reject classify ambiguous samples with level SVM classifier. A non-symmetric thresholding scheme applied: different thresholds defined for positive negative samples. The rejected forwarded second LR Finally choose way combine decisions obtain final rule. experiments performed UCI Arrhythmia Database.