摘要
为了实现便携式近红外光谱仪水果糖度现场快速分析,将桃、梨和苹果的光谱进行二阶导数和卷积平滑处理后,利用组合移动窗口偏最小二乘法选择信息变量建立PLS模型,利用遗传偏最小二乘法选择信息变量建立MLR模型。分析表明,桃、梨和苹果PLS模型的RMSEP分别为0.417、0.372和0.654,其RSDP分别为4.685%、3.348%和4.111%;MLR模型的RMSEP分别为0.381、0.382和0.550,其RSDP分别为4.281%、3.438%和3.457%,模型预测精度均满足现场检测应用要求。结果表明,用SCM-WPLS和GA-PLS可以提取最有效信息变量,模型更加简洁、数据运算量也更少,模型适用于便携式近红外光谱仪器。
To non-destructive determinate the soluble solids content(SSC) of peach,pear and apple using portable Vis/NIR instrument,searching combination moving window partial least squares(SCMWPLS) and genetic algorithm partial least squares(GA-PLS) are proposed to select the informative variables,partial least squares(PLS) and multiple linear regression(MLR) models are developed,respectively.The optimal PLS models of peach,pear and apple were obtained with the root mean square error of prediction(RMSEP) of 0.417,0.372 and 0.654 and the relative standard deviation of prediction(RSDP) of 4.685%,3.348% and 4.111%,respectively;the optimal MLR models of peach,pear and apple were obtained with the RMSEP of 0.381,0.382 and 0.550 and the RSD P of 4.281%,3.438% and 3.457%,respectively.It shows that the SCMWPLS and GA-PLS procedures are very promising for vibrational spectroscopy based on multivariate analysis and give approving prediction precision for the determination of SSC in fruits by portable Vis/NIR instrument.
出处
《食品安全质量检测学报》
CAS
2009年第1期32-38,共7页
Journal of Food Safety and Quality
基金
国家自然科学基金(30571073)
"十一五"国家科技支撑计划(2006BAD05A06-Z012)资助项目
关键词
便携式近红外光谱仪
变量选择
偏最小二乘法
多元线性回归
水果
糖度
Portable Vis/NIR instrument
Variable selection
partial least squares
multiple linear regression
Fruits,Soluble solids content