摘要
选取150份干扰物和浓度不同的葡萄糖样本数据,按照全干扰、缺失胆固醇、缺失乳酸、缺失白蛋白和缺失尿素将数据划分5个子集。每个子集进行卷积平滑滤波法(Savitzky-Golay smoothing,SG)处理后,建立偏最小二乘回归(Partial Least Squares Regression,PLSR)模型。利用克拉克误差网格(Clarke Error Grid,CEG)及t检验分析4种干扰物对葡萄糖预测影响。结果表明,子集1~5对应模型预测集相关系数(Correlation Coefficient of Prediction,R p)分别为0.9131、0.7115、0.7624、0.8578和0.8658,预测集均方根误差(Root Mean Square Error of Prediction,RMSEP)分别为54.8993、239.5512、162.3715、133.9682和106.0521 mg/dL。5个子集位于CEG的A+B区分别为100%、71.43%、66.66%、85.71%和88.89%。t检验中每1 mg/dL的胆固醇、乳酸和白蛋白分别使葡萄糖预测值降低5.288 mg/dL、增高2.214 mg/dL和增高0.031 mg/dL。故胆固醇和乳酸的影响相当显著,其次是白蛋白,而尿素的影响则相对较弱。因此,在中红外血糖定量分析中必须考虑胆固醇、乳酸和白蛋白对血糖检测的影响。
Taking 150 parts of glucose sample data containing different interferences and different concentrations,the infrared spectral data were divided into 5 subsets according to full interference,missing cholesterol,missing lactate,missing albumin and missing urea.Each subset was smoothed by Savitzky-Golay smoothing(SG)to construct a partial least squares regression(PLSR)model.The Clarke Error Grid(CEG)and t-test were used to analyze the influence of four interferents on glucose prediction.The results showed that the corresponding model correlation coefficient of prediction(R p)for subsets 1 to 5 were 0.9131,0.7115,0.7624,0.8578 and 0.8658,respectively,and the root mean square error of prediction(RMSEP)were 54.8993,239.5512,162.3715,133.9682 and 106.0521 mg/dL.The five subsets fell in the A+B zone of CEG,which were 100%,71.43%,66.66%,85.71%and 88.89%,respectively.Cholesterol(decreased by 5.288 mg/dL),lactate(increased by 2.214 mg/dL)and albumin(increased by 0.031 mg/dL)per 1 mg/dL in t-test decreased and increased the predictive value of glucose,respectively.So,cholesterol and lactate have significant effects,followed by albumin,urea has relatively weak effects.Therefore,the effects of cholesterol,lactate and albumin on blood glucose detection must be considered in the quantitative analysis of mid-infrared blood glucose.
作者
吕亚玲
张朱珊莹
冉康
岳岩松
张献文
曹汇敏
LV Ya-ling;ZHANG Zhu-shan-ying;RAN Kang;YUE Yan-song;ZHANG Xian-wen;CAO Hui-min(College of Biomedical Engineering,South-Central MinZu University,Wuhan 430074,China;Key Laboratory of Cognitive Science,State Ethnic Affairs Commission,Wuhan 430074,China;Hubei Key Laboratory of Medical Information Analysis and Tumor Diagnosis&Treatment,Wuhan 430074,China;Linyi Grepo Garden Machinery Co,Ltd,Linyi 276700,China)
出处
《化学试剂》
CAS
北大核心
2023年第10期14-20,共7页
Chemical Reagents
基金
国家自然科学基金项目(61501526,61178087)
中南民族大学省级创新创业训练项目(SCX2023068)。
关键词
中红外光谱
光谱分析
血糖检测
干扰物
定量模型
mid-infrared spectrum
spectral analysis
blood glucose detection
interferents
quantitative models