Discrimination of Blended Chinese Rice Wine Ages Based on Near-Infrared Spectroscopy

Abstract

In this study, near-infrared spectroscopy combined with spectral preprocessing methods was used for the discrimination of blended Chinese rice wine ages (3, 5, 8, and 10 years aged). Discriminant models were developed using principal component analysis, linear discriminant analysis, and discriminant partial least squares regression. The correct classifications for young wines (3 and 5 years) and aged wines (8 and 10 years) were 100% using discriminant partial least squares after spectral preprocessing. Moreover, for the classification of rice wines from the four years aged groups, 95.0% classification accuracy was obtained using discriminant partial least squares with orthogonal signal correction pretreatment in a validation sample set.