Department of Sciences and Technological Innovation, University of Eastern Piedmont, Viale Michel 11, 15121 Alessandria, Italy.
Department of Pharmaceutical and Toxicological Chemistry, University of Napoli Federico II, Via Montesano 49, 80131 Naples, Italy.
Spectrochim Acta A Mol Biomol Spectrosc. 2018 Jan 15;189:427-435. doi: 10.1016/j.saa.2017.08.050. Epub 2017 Aug 17.
The authentication and traceability of hazelnuts is very important for both the consumer and the food industry, to safeguard the protected varieties and the food quality. This study investigates the use of a portable FTIR spectrometer coupled to multivariate statistical analysis for the classification of raw hazelnuts. The method discriminates hazelnuts from different origins/cultivars based on differences of the signal intensities of their IR spectra. The multivariate classification methods, namely principal component analysis (PCA) followed by linear discriminant analysis (LDA) and partial least square discriminant analysis (PLS-DA), with or without variable selection, allowed a very good discrimination among the groups, with PLS-DA coupled to variable selection providing the best results. Due to the fast analysis, high sensitivity, simplicity and no sample preparation, the proposed analytical methodology could be successfully used to verify the cultivar of hazelnuts, and the analysis can be performed quickly and directly on site.
榛子的身份验证和可追溯性对消费者和食品行业都非常重要，以保护受保护的品种和食品质量。本研究调查了使用便携式傅里叶变换红外（FTIR）光谱仪结合多元统计分析对生榛子进行分类的方法。该方法基于其红外光谱信号强度的差异来区分来自不同产地/品种的榛子。多元分类方法，即主成分分析（PCA）后接线性判别分析（LDA）和偏最小二乘判别分析（PLS-DA），无论是否进行变量选择，都可以很好地对组进行区分，而结合变量选择的 PLS-DA 提供了最佳结果。由于分析速度快、灵敏度高、简单且无需样品制备，因此所提出的分析方法可以成功用于验证榛子的品种，并且可以快速直接地在现场进行分析。
