Liu Yan-De, Ying Yi-Bin
College of Biosystems Engineering and Food Science, Zhejiang University, Hangzhou 310029, China.
Guang Pu Xue Yu Guang Pu Fen Xi. 2006 Aug;26(8):1454-6.
Fourier transform near infrared(FT-NIR) spectrum of intact Xueqing pear was obtained by diffusion reflectance. Partial least squares (PLS) regression was carried out, describing the relationships between the data sets of laboratory data and the FT-NIR spectra. Different wave number ranges were chosen for regression and spectral information abstraction. The 3D-curves were shown for different factors, root mean square errors of cross validation (RMSECV), and prediction residual sum of squares (PRESS). Analysis results show that the best calibration model gave the relative high correlation coefficient of 0. 79 and the low standard errors of prediction of 0. 019 when the best wave number range was 5 452-12 285 cm(-1) and the best factor was 7. The method of selecting advantageous wavelength ranges is feasible to obtain high prediction precision.
采用漫反射法获取了完整雪青梨的傅里叶变换近红外（FT-NIR）光谱。进行了偏最小二乘（PLS）回归，描述了实验室数据与FT-NIR光谱数据集之间的关系。选择不同的波数范围进行回归和光谱信息提取。给出了不同因子、交叉验证均方根误差（RMSECV）和预测残差平方和（PRESS）的三维曲线。分析结果表明，当最佳波数范围为5452 - 12285 cm⁻¹且最佳因子为7时，最佳校准模型给出了相对较高的相关系数0.79和较低的预测标准误差0.019。选择有利波长范围的方法对于获得高预测精度是可行的。
