面向呼吸腰带数据的病症正常异常分类  

Respiratory Belt Data Oriented Normal and Abnormal Classification

作　　者：徐正丽 蒋盟珂 谢梅英 蔡翔[1] XU Zhengli;JIANG Mengke;XIE Meiying;CAI Xiang(Guilin University of Electronic Technology,Guilin,Guangxi,541004,China;Nanjing University of Information Science and Technology,Nanjing,Jiangsu,210044,China)

机构地区：[1]桂林电子科技大学,广西桂林541004 [2]南京信息工程大学,江苏南京210044

出　　处：《广西科学》2022年第2期241-248,共8页Guangxi Sciences

基　　金：国家自然科学基金项目(71463010,22180155466);财政部科技专项(52544788);广西自然科学基金项目(2014GX233658);江苏省研究生科研创新计划项目(KYCX21_1040)资助。

摘　　要：人体的呼吸信号包含了很多指示呼吸健康的信息,虽然有多种测量呼吸的仪器和手段,但是呼吸腰带仍然是经济、没有辐射伤害、能长时程日常佩戴的呼吸数据获取手段。然而,目前呼吸腰带数据的分析方法还不成熟,为了明确哪些呼吸数据特征可以对病理进行有效分类,本研究分别使用长短期记忆网络(Long Short-Term Memory,LSTM)、吸呼比结合LSTM、吸呼比结合支持向量机(Support Vector Machine,SVM)3种方法,对287个正常呼吸和55个异常呼吸的24 h观测数据进行病理类别分类准确度对比实验。结果表明,吸呼比特征结合LSTM法具有更高的分类准确度。The respiratory signals of human body contain a lot of information indicating respiratory health.Although there are a variety of instruments and means to measure respiration,the respiratory belt is still an economic,no radiation damage,and long-term daily wearable means to obtain respiratory data.However,the current analysis method of respiratory belt data is not mature.In order to clarify which respiratory data characteristics can effectively classify pathology,this study used three methods,namely Long Short-Term Memory(LSTM),inspiratory/expiratory ratio combined with LSTM,and inspiratory/expiratory ratio combined with Support Vector Machine(SVM),to conduct a comparative experiment on the accuracy of pathological classification of 24 h observation data of 287 normal respirations and 55 abnormal respirations.The results showed that the inspiratory/expiratory ratio feature combined with the LSTM method has higher classification accuracy.

关 键 词：呼吸腰带 吸呼比 特征分析 长短期记忆网络 支持向量机 

分 类 号：TN911.6[电子电信—通信与信息系统]