净信号的局部建模算法及其在近红外光谱分析中的应用  

Local Regression Algorithm Based on Net Analyte Signal and Its Application in Near Infrared Spectral Analysis

作　　者：张红光[1] 卢建刚[1] 

机构地区：[1]浙江大学工业控制技术国家重点实验室,浙江杭州310027

出　　处：《光谱学与光谱分析》2016年第2期384-387,共4页Spectroscopy and Spectral Analysis

基　　金：国家(973计划)项目(2012CB720500);国家自然科学基金项目(61590925)资助

摘　　要：提出了一种基于净信号分析的局部建模算法,以克服光谱定量分析中样本间差异性过大和样本待测性质与光谱之间存在非线性等问题。首先利用净信号分析方法得到校正样本和待测样本的净信号,然后用待测样本净信号和校正样本净信号之间的欧式距离作为样本相似性判据,选取一定数量的与待测样本最相似的校正样本组成局部校正子集,建立局部PLS回归模型。针对一组猪肉近红外光谱数据集的实验结果表明,该方法的预测精度显著优于全局建模方法和基于光谱欧式距离的局部建模方法。To overcome the problems of significant difference among samples and nonlinearity between the property and spectra of samples in spectral quantitative analysis,a local regression algorithm is proposed in this paper.In this algorithm,net signal analysis method（NAS）was firstly used to obtain the net analyte signal of the calibration samples and unknown samples,then the Euclidean distance between net analyte signal of the sample and net analyte signal of calibration samples was calculated and utilized as similarity index.According to the defined similarity index,the local calibration sets were individually selected for each unknown sample.Finally,a local PLS regression model was built on each local calibration sets for each unknown sample.The proposed method was applied to a set of near infrared spectra of meat samples.The results demonstrate that the prediction precision and model complexity of the proposed method are superior to global PLS regression method and conventional local regression algorithm based on spectral Euclidean distance.

关 键 词：光谱定量分析 局部校正方法 净信号 偏最小二乘 近红外光谱 

分 类 号：O657.3[理学—分析化学]