【摘要】 目前,人们对房颤维持和终止的机制还没有完全了解,因此对阵发性房颤和持续性房颤的分类具有非常重要的研究意义。鉴于此,本研究提出一种新的分类方法。根据主成分分析从单导联心电信号中提取出房颤信号,其次计算提取到的房颤信号的特征,最后用分类器对阵发性和持续性房颤进行分类。提出将房颤波的复杂度作为房颤波波动复杂度的表征。对阵发性和持续性房颤分类的实验结果表明,预测的总正确率是90%。在1 000次随机性实验中,最高分类正确率可达到92%,平均正确率为77.12%。该方法可以很好的对两类房颤进行分类,对预测房颤的自发性终止有一定的指导意义。更多还原

【Abstract】 Up to date the mechanisms of maintenance and termination of atrial fibrillation(AF) have not been understood adequately;therefore it is very important research work to classify paroxysmal and sustained atrial fibrillations.In this paper,a novel method for classifying paroxysmal and sustained atrial fibrillations was proposed.The method first extracted AF signal from single-lead electrocardiogram(ECG) based on principal component analysis;then calculated the feature parameters of AF signals;finally classified the paroxysmal and sustained atrial fibrillations using classier.Results showed that general accuracy rate of the prediction was 90%,and for the 1,000 random prediction experiments,the maximum accurate was 92%,with an average accuracy of 77.12%,which proved that the method performed well in classification of paroxysmal and sustained atrial fibrillations.The proposed method showed potentials of serving as guidance in predicting the spontaneous termination of AF.更多还原