Application of multiblock data analysis for enhancing the predictive performance of sugar contents using near infrared spectrometry

Abstract

Due to the benefits of various detection modes in the near-infrared (NIR) spectroscopic technique, this research combined the NIR spectral data from the different detection modes aiming to enhance the quantification performance of the instrument. The mixing sugar samples, composed of glucose, fructose, maltose, and sucrose, and established based on mixture designs, were used for the demonstration. The NIR spectra were recorded from portable Vis-NIR and benchtop NIR spectrometers using both transmittance and transflectance detection modes. Multiblock-principal component analysis (MB-PCA) was applied to exploratorily analyze the multiblock data. Multiblock regression models, including concatenated-partial least squares (C-PLS), serial-PLS (S-PLS), and multiblock-PLS (MB-PLS), were employed to quantify the concentration of the sugar samples. The MB-PCA could effectively identify the difference among the studies sugars. In addition, the spectral fusion using the multiblock data analysis could improve the predictive performance in terms of the R2, Q2, root mean square error of calibration (RMSEC), root mean square error of cross validation (RMSECV), ratio of prediction to deviation (RPD), and relative standard deviation (RSD) values when compared with the conventional PLS.