基于小波的数据挖掘技术在Holter心电信号分析中的应用  

Application of Holter ECG Signal Analysis Based on Wavelet and Data Mining Technique

作　　者：余辉[1] 张力新[1] 刘文耀[1] 黄志勇[1] 丁明石[1] 

机构地区：[1]天津大学精密仪器与光电子工程学院,天津300072

出　　处：《天津大学学报》2006年第B06期153-156,共4页Journal of Tianjin University(Science and Technology)

基　　金：天津市重点自然科学基金(003609111).

摘　　要：针对数据挖掘技术的应用,分析了小波变换的多分辨率特性和带通滤波的本质．按照数据挖掘的思想提出基于二次微分小波概貌信号的Holter心电信号中R波极值点检测的数学模型,采用Mallat递归滤波的算法计算离散小波变换并用等效滤波器的思想对算法进行优化．该检测算法经MIT/BIH心律失常数据库的检验,R波有效检测率高达99%；在100多例临床测试中检测率也高达97%以上．该算法模块现已经应用到商业软件中并取得令人满意的效果．The paper introduces the application of data mining technique and analyze the multi-frequency and strap..pass filter essence of wavelet transform in detail. Based on data mining theory and dyadic differential wavelet, new model for R peak detection in Holter ECG signal is proposed. We use the Mallat recursive filter algorithm to calculate wavelet transform and optimize the detection algorithm based on the equivalent filter technique. The detection algorithm is tested with MIT/BIH arrhythmia database and reached the efficiency of 99% for R peak detection. In over 100 subjects in hospital applied with this algorithm ,the test result reached 97% effective detection rate. Now this algorithm module has been applied in business software and received great satisfaction from the users.

关 键 词：小波变换 数据挖掘 信号检测 心电图 二进小波 R波检测 

分 类 号：R318.04[医药卫生—生物医学工程]