Abstract

Partial least squares (PLS) is a powerful tool for multivariate linear regression. But what if the data show a non-linear structure? Near infrared spectra from a pharmaceutical process were used as a case study. An ANOVA test revealed that the data are well described by a 2^nd order polynomial. This work investigates the application of regression techniques that account for slightly non-linear data. The regression techniques investigated are: linearising data by applying transformations, local PLS, i.e. splitting of data, and quadratic PLS. These models were compared with ordinary PLS and principal component regression (PCR). The predictive ability of the models was tested on an independent data set acquired a year later. Using the knowledge of non-linear pattern and important spectral regions, simpler models with better predictive ability can be obtained.

© 2003 NIR Publications

Linear regression models and k-means clustering for statistical analysis of fNIRS data

Viola Bonomini, Lucia Zucchelli, Rebecca Re, Francesca Ieva, Lorenzo Spinelli, Davide Contini, Anna Paganoni, and Alessandro Torricelli
Biomed. Opt. Express 6(2) 615-630 (2015)

EGN model of non-linear fiber propagation

Andrea Carena, Gabriella Bosco, Vittorio Curri, Yanchao Jiang, Pierluigi Poggiolini, and Fabrizio Forghieri
Opt. Express 22(13) 16335-16362 (2014)

Abel inversion of deflectometric data: comparison of accuracy and noise propagation of existing techniques

Pankaj S. Kolhe and Ajay K. Agrawal
Appl. Opt. 48(20) 3894-3902 (2009)

Quantitative analysis of bayberry juice acidity based on visible and near-infrared spectroscopy

Yongni Shao, Yong He, and Jingyuan Mao
Appl. Opt. 46(25) 6391-6396 (2007)

Discussion on the validity of NIR spectral data in non-invasive blood glucose sensing

Wanjie Zhang, Rong Liu, Wen Zhang, Hao Jia, and Kexin Xu
Biomed. Opt. Express 4(6) 789-802 (2013)