Authentication of apple juice categories based on multivariate analysis of the synchronous fluorescence spectra

Partial least squares discriminant analysis (PLS-DA) was used for the development of the classification models. The analysis was performed for the unfolded TSFS (uTSFS) and individual synchronous fluorescence spectra (SFS) measured at particular emission-excitation wavelength offsets (Δλ), from 10 to 160 nm with 10 nm step. The best discrimination results were obtained for the model using uTSFS in the range of Δλ = 30-160 nm, with cross-validation and external validation error rates of 0.05 and 0.06, respectively. Models based on selected individual SFS also showed similarly good predictive ability, with cross-validation and external validation error rates of 0.08 and 0.05, respectively, for the SFS measured at Δλ = 70 and 90 nm. The analysis of significant variables using selectivity ratio (SR) suggests that the fluorescence of non-enzymatic browning products may significantly contribute to the differentiation of FC and NFC juices.