Abstract

Info

Title:: ECG-based multi-level conjugate symmetric Hadamard feature transformation for classification of abnormal signals of atrial fibrillation

Author(s):: WANG Kai; YANG Shu; LI Chao; Department of Health Management, Bengbu Medical College, Bengbu 233030, China

Keywords:: ; Keywords: atrial fibrillation; electrocardiogram; multi-level conjugate symmetric Hadamard feature transformation; Levenberg-Marquardt neural network

Abstract:: Abstract: Objective To detect and analyze the abnormal signals of atrial fibrillation (AF) by multi-level conjugate symmetric Hadamard feature transformation model for establishing a system to classify AF abnormal signals. Methods The features for the classification of AF abnormal signals were detected with multi-level conjugate symmetric Hadamard feature transformation. Levenberg-Marquardt neural network model based on error gradient back-propagation was used for the training of test data set. A classifier for the classification of AF abnormal signals was constructed, and finally a classification model used in clinical diagnosis was established. Results The proposed model effectively improved the performance of feature classification, increased convergence speed and algorithm accuracy, thereby facilitating the real-time analysis and diagnosis of AF. Conclusion The proposed model which has high system robustness can be used to capture suspected waveforms of AF abnormal signals, evaluate and analyze signal features.