摘要
比较了偏最小二乘(PLS)法在常规和一阶导数近红外漫反射光谱法分析安乃近的定量预测能力;讨论了校正集样品数和波长间隔对PLS预测的影响.在常规光谱法中,校正样品和预测样品的预测结果的相对标准误差分别为0.9891%和1.4630%.而在一阶导数光谱法中,校正样品和预测样品的预测结果的相对标准误差分别为0.1102%和1.0429%.
The solid samples of analgini were analyzed noninvasively by means of conventional and first derivative near-infrared diffuse reflectance spectra. The quantitative prediction abilities of partial least squares methods in both cases were compared. The influence of the calibration samples and wavelengths on the Capacity of PLS are discussed. The relative standard errors of prediction for the calibration samples and prediction samples with conventional spectra are 0.9891% and 1. 463 0%,while those with derivative spectra are 0. 110 2% and 1. 042 9%, respectively.
出处
《吉林大学自然科学学报》
CAS
CSCD
1997年第3期99-102,共4页
Acta Scientiarum Naturalium Universitatis Jilinensis
关键词
偏最小二乘法
安乃近
非破坏分析
近红外光谱
partial least squares method,near-infrared diffuse reflectance spectraum,analgini,noninvasive analysis