基于改进K均值聚类生成匹配模板的心搏分类方法

ECG beat classification method based on match template generated by improved K-means clustering

为提高分析含大量数据的动态心电时的准确性和分析效率,提出了一种基于改进的K均值聚类生成心搏模板的匹配方法。使用K均值聚类和波形反混淆技术进行循环纠错,生成可变宽心搏模板、并建立心搏模板库。利用可变宽心搏模板和相关系数相结合的策略,对动态心电中心搏进行快速准确分类。实验方法经心率失常数据库MIT-BIT和ANMA/ANSI标准验证,分类结果总体准确率达98.06%,达到了心搏分类目标。

To improve accuracy and efficiency when analyze morphology of large dataset of dynamic electrocardiography（ECG）,a ECG beat classification method based on beat templates generated by improved Kmeans clustering is presented. It takes K-means clustering and DEMIX technology to correct errors circularly,generates variable-width beat templates and establish beat templates database. By using the strategy which combines variable-width beat templates with correlation coefficient,the method can classify the ECG beats efficiently and accurately. The experimental verification is evaluated on the MIT-BIT arrhythmia database and ANMA/ANSI standard,and the overall accuracy rate of classification result is 98. 06 %,it achieves the goal of beats classification.