基于小波变换和Adaboost算法的心脏骤停预测模型研究  

The Study on The Sudden Cardiac Arrest Prediction Model based on Wavelt Tansform and Adaboost Algorithm

作　　者：刘光达[1] 王永祥[1] 蔡靖[1] 王伟[1] 刘忠民[1] 

机构地区：[1]吉林大学仪器科学与电气工程学院,长春130061

出　　处：《生物医学工程研究》2017年第2期95-100,共6页Journal Of Biomedical Engineering Research

基　　金：国家"十二五"科技支撑计划课题(2015BAI02B04);吉林省教育厅"十二五"科学技术研究规划重点项目(440020031134);吉林省发改委省级产业创新专项资金项目(2016C052-2);吉林市科技发展计划项目(2015313013)

摘　　要：为了预测心脏骤停,应用小波变换和Adaboost算法建立心脏骤停预测模型。首先用小波变换方法对正常窦性心律心电数据和有心脏骤停症状患者的心电数据进行分析、提取特征值,再用Adaboost算法对两种数据进行分类来预测心脏骤停的发生。实验验证,本模型分类预测效果较好,在心脏骤停发生前5min,其预测精度高达97.56%,为心脏骤停的预测提供了一种有参考价值的方法。In order to predict cardiac arrest,wavelet transform and adaboost algorithm were used to establish the prediction model of cardiac arrest. First,wavelet transform was used to analyze and extract the characteristics value of normal sinus rhythm ECG and sudden cardiac arrest（ SCA） ECG. Then the two datas were classified by adaboost to predict the incidence of SCA. Experiments verify that this model is effective in SCA prediction in the 5 min before SCA occurs,its prediction accuracy is 97. 56%.It provides a valuable method for the prediction of cardiac arrest.

关 键 词：心脏骤停 小波变换 ADABOOST 预测 特征选择 

分 类 号：R318[医药卫生—生物医学工程]