Aiming at the problem of low diagnostic accuracy and positive predictive value of imbalanced datasets in arrhythmia diagnosis algorithms,an auxiliary method for arrhythmia diagnosis based on optimized adaptive models was proposed.This method extracted the 77 dimensional features of the electrocardiogram signal and fuses them,trained the diagnostic model using the fused features,and optimized the adaptive model parameters using an improved particle swarm optimization algorithm.The optimized model was used to test in MIT-BIH arrhythmia database and compared with the existing methods.The results show that the total diagnostic accuracy of the proposed method on the test dataset reaches 98.2％,and the positive predictive values of normal or bundle branch block rhythm,supraventricular abnormal rhythm,ventricular abnormal rhythm,and fusion rhythm reach 98.5％,96.1％,95.5％,and 92.0％,respectively.The diagnostic accuracy and positive predictive value are significantly higher than those of the existing methods.
维普
万方数据