基于电子鼻的2型糖尿病虚实病性间的气味识别分析  

Recognition and analysis on smell between deficiency and excess syndromes of patients with type 2 diabetes mellitus based on electronic nose

作　　者：林雪娟[1,2] 郑哲洲[1,2] 吴青海[1,2] 刘丽桑[3] 郑剑翔[4] 李灿东[1,2] 

机构地区：[1]福建中医药大学中医证研究基地,福州350122 [2]福建省中医健康状态辨识重点实验室(福建中医药大学),福州350122 [3]福建工程学院,福州350118 [4]福州大学,福州350122

出　　处：《中华中医药杂志》2015年第8期2687-2691,共5页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家自然科学基金项目(No.81373552);福建省自然科学基金项目(No.2014J01362);福建省发改委产业技术开发项目;福建省中医药科研项目(No.wzpw201313)~~

摘　　要：目的:运用电子鼻技术探讨2型糖尿病(T2DM)患者及其虚实病性间的口腔呼气的气味识别。方法:选择200例T2DM患者和100名健康者,运用基于薄膜型气体传感器阵列技术的医用电子鼻(EN0l1103-A)采集口腔呼气的气味图谱,采用支持向量分类机的方法识别T2DM患者和健康者的气味图谱特征,同时选择气味图谱响应曲线的振幅和斜率作为图谱特征参数,比较和识别T2DM虚实病性间的气味图谱特征。结果:1采用支持向量分类机的方法,选择高斯核对T2DM患者和健康者的气味图谱特征的识别准确率高达88.04%。2T2DM患者实证组的气味图谱响应曲线D的振幅高于虚实夹杂组(P<0.05);实证组的气味图谱响应曲线D的斜率显著高于虚证组和虚实夹杂组(P<0.01),虚证组和虚实夹杂组的气味图谱响应曲线G的斜率显著高于实证组(P<0.05,P<0.01)。3在响应曲线G的斜率特征上,采用Fisher判别分析对T2DM实证组、虚证组及虚实夹杂组的气味图谱特征的识别准确率分别为75.0%、74.6%、52.9%,对全部案例的识别准确率为62.4%。结论:运用电子鼻技术可以较准确地识别T2DM患者口腔呼气的气味特征,并能初步判断T2DM的虚实病性。电子鼻技术可以为中医嗅诊客观化研究提供新的手段和方法。Objective： To explore the identification methods for the smell of mouth breathing between deficiency and excess syndromes of patients with type 2 diabetes mellitus （T2DM） by using electronic nose. Methods： 200 T2DM patients and 100 healthy people were included, and the characteristic fingerprint of mouth breathing smell was gathered by using electronic nose （EN011103-A） supporting by a thin film type of gas sensor array technology, and identified by using a classification method based on supporting vector classification. The amplitude and slope of smell fingerprint response curve were selected as the fingerprint characteristic parameter in order to compare and identify the smell fingerprint characteristics between deficiency and excess syndromes of T2DM patients. Results： ①The accuracy rate of using classification method based on supporting vector classification and selecting Ganssian kernel to identify the smell fingerprint characteristics of T2DM patients and healthy people was as much as 88.04%. ②The amplitude of smell fingerprint response curve D of T2DM patients with excess syndrome was higher than that of T2DM patients with syndrome of intermingled deficiency and excess （P〈0.05）, the slope of smell fingerprint response curve D of T2DM patients with excess syndrome was significantly higher than that of T2DM patients with intermingled deficiency and excess syndrome and deficiency syndrome （P〈0.01）, and the slope of smell fingerprint response curve G of T2DM patients with intermingled deficiency and excess syndrome and deficiency syndrome was significantly higher than that of T2DM patients with excess syndrome （P〈0.05, P〈0.01）. ③The slope characteristic of smell fingerprint response curve G showed that the accuracy rate of using fisher discriminant analysis model to identify the smell fingerprint characteristic of T2DM patients with excess syndrome, deficiency syndrome and intermingled deficiency and excess syndrome was 75.0%, 74.6% and 52.9% respectively, and the identi

关 键 词：嗅诊 电子鼻 2型糖尿病 病性 虚实 

分 类 号：R259[医药卫生—中西医结合]