Liang, Z.Z. and Shi, P.F. (2004) An Efficient and Effective Method to Solve Kernel Fisher Discriminant Analysis. Neurocomputing, 61, 485-493.
基于近红外光谱技术快速鉴别白酒真伪Identification of Specific Liquor Based on Near-Infrared Spectroscopy Technology
近红外光谱, 白酒, 鉴别, Fisher判别法Near-Infrared Spectroscopy, Liquor, Identification, Fisher Linear Discriminant
《Advances in Analytical Chemistry》, Vol.6 No.1, 2016-02-05
本文探索了一种利用近红外光谱技术进行白酒真伪快速鉴别的新方法。共收集到分属8个品牌的100个酒样，基于5340~7400 cm−1的光谱区域，使用四种化学计量学方法建立分类模型，即基于全部变量的K最近邻模型、基于前四个主成分得分的K最近邻模型，基于前四个主成分得分的感知机、基于全部变量的Fisher判别模型。比较显示，Fisher判别模型达到了最优的性能，在训练集和测试集上的误判率均为0%。可见近红外光谱技术结合模式识别方法可用于快速鉴别五粮液等高端酒真伪。A rapid nondestructive method for kind identification of liquor by near infraredspectroscopy (NIR) was investigated and developed. A total of 100 samples belonging to 8 brands were collected in local markets. Based on the wavenumber region of 5340-7400 cm−1, four kinds of algorithm, i.e., K-nearest neighbor’s method of all spectral variables (KNN1) and of the first four principal components (KNN2), perception of the first four principal components and Fisher linear discrimination of all spectral variables, were used for constructing the classification models. By comparison, it shows that the Fisher linear discrimination model can achieve the best performance, with a misclassified ratio of 0%in either the training set or the test set. The results also indicate that the combination of NIR and pattern recognition models is feasible for a rapid identification of liquor such as Wuliangye.