Cavdaroglu Cagri, Ozen Banu
Department of Food Engineering, Izmir Institute of Technology, 35430 İzmir, Türkiye.
Foods. 2023 Mar 8;12(6):1139. doi: 10.3390/foods12061139.
Spectroscopic techniques as untargeted methods have great potential in food authentication studies, and the evaluation of spectroscopic data with chemometric methods can provide accurate predictions of adulteration even for hard-to-identify cases such as the mixing of vinegar with adulterants having a very similar chemical nature. In this study, we aimed to compare the performances of three spectroscopic methods (fluorescence, UV-visible, mid-infrared) in the detection of acetic-acid/apple-vinegar and spirit-vinegar/apple-vinegar mixtures (1-50%). Data obtained with the three spectroscopic techniques were used in the generation of classification models with partial least square discriminant analysis (PLS-DA) and orthogonal partial least square discriminant analysis (OPLS-DA) to differentiate authentic and mixed samples. An improved classification approach was used in choosing the best models through a number of calibration and validation sets. Only the mid-infrared data provided robust and accurate classification models with a high classification rate (up to 96%), sensitivity (1) and specificity (up to 0.96) for the differentiation of the adulterated samples from authentic apple vinegars. Therefore, it was concluded that mid-infrared spectroscopy is a useful tool for the rapid authentication of apple vinegars and it is essential to test classification models with different datasets to obtain a robust model.
光谱技术作为非靶向方法在食品真伪鉴定研究中具有巨大潜力，并且使用化学计量学方法对光谱数据进行评估，即使对于难以识别的情况，如将醋与化学性质非常相似的掺假物混合，也能提供掺假的准确预测。在本研究中，我们旨在比较三种光谱方法（荧光、紫外可见、中红外）在检测醋酸/苹果醋和蒸馏醋/苹果醋混合物（1%-50%）方面的性能。使用这三种光谱技术获得的数据用于通过偏最小二乘判别分析（PLS-DA）和正交偏最小二乘判别分析（OPLS-DA）生成分类模型，以区分正品和混合样品。通过一些校准集和验证集，采用一种改进的分类方法来选择最佳模型。只有中红外数据提供了强大而准确的分类模型，对掺假样品与正品苹果醋的区分具有高分类率（高达96%）、灵敏度（1）和特异性（高达0.96）。因此，得出结论，中红外光谱是苹果醋快速真伪鉴定的有用工具，并且使用不同数据集测试分类模型以获得稳健模型至关重要。
