【摘要】 目的:探讨基于多尺度快速样本熵与随机森林的心电图分析方法对常见心律失常(房性早搏、室性早搏)的自动诊断的可行性和有效性。方法:利用不同心律失常疾病的心电信号存在复杂性差异的特点,通过多尺度熵计算心电信号在不同尺度下的样本熵值以组成特征向量;利用kd树提高多尺度熵的计算效率,增强算法的实时性。利用训练样本的特征向量构建随机森林分类器,再根据众多决策树的分类结果结合投票原则确定测试样本心律失常疾病的类型。结果:本文提出的心电图分析方法能够有效地识别正常心律、房性早搏(APB)及室性早搏(VPB),平均识别准确率达到91.60%。结论:本文提出的心电图分析方法对常见心律失常(APB,VPB)具有较高的识别准确率及临床实用价值。更多还原

【Abstract】 Objective:To explore the feasibility and effectiveness of ECG analysis method based on multiscale fast sample entropy and random forest for automatic diagnosis of common arrhythmia(atrial premature beat,ventricular premature beat).Methods:Different arrhythmia diseases having the features of complexity difference of ECG signals are adopted,and the sample entropy of ECG signal at different scales calculated by the multi-scale entropy forms eigenvectors;Using kd tree to improve the computation efficiency of multiscale entropy,the real-time performance of algorithm is enhanced.The random forest classifier is constructed by the eigenvectors of training samples,and then the type of arrhythmias is determined by the classification results of the numerous decision trees coupled with voting principle.Results:The proposed electrocardiogram analysis method can effectively identify the normal heart rhythm,atrial premature beats(APB) and ventricular premature beats(VPB),with an average identification accuracy of 91.60%.Conclusions:The ECG analysis method presented in this paper has high recognition accuracy and clinical value for common arrhythmia(APB,VPB).更多还原