Using iSPA-PLS and NIR spectroscopy for the determination of total polyphenols and moisture in commercial tea samples
In this work, a methodology is proposed for determining the content of total polyphenols and moisture in commercial tea samples by using near-infrared spectroscopy (NIRS) and Partial Least Squares (PLS) regression coupled with the Successive Projections Algorithm for interval selection (iSPA-PLS). For comparison, full-spectrum PLS and the Interval PLS (iPLS) were also used. Since the spectra are scattered and exhibit systematic variations on the baseline, standard normal variate transformation (SNV) and multiplicative scatter correction (MSC) were applied as data preprocessing methods. The number of PLS latent variables and the number of region intervals were optimized according to the root mean square error of cross-validation (RMSECV) and coefficient of determination (RCV2) in the calibration set. The predictive ability of the final model was evaluated in terms of the root mean square error of prediction (RMSEP), coefficient of determination (RPred2) and ratio performance deviation (RPDPred) in the external prediction set, which were not employed in the model-building process. For the determination of the total polyphenol content, 10-iSPA-PLS with MSC preprocessing presented the best results with the smallest RMSEP (0.599 mg kg−1), and the highest RPred2 (0.933) and RPDPred (3.863) values. For the determination of moisture content, 20-iSPA-PLS with MSC preprocessing achieved the best results with the smallest RMSEP (0.32 mg kg−1), and the highest RPred2 (0.94) and RPDPred (4.08) values. Thus, it can be concluded that the NIRS coupled with iSPA-PLS is a promising analytical tool for monitoring tea quality.