关键词: 羊奶, 蛋白质掺假, 电子鼻, 化学计量法

Abstract: Goat milk adulterated with proteins was qualitatively discriminated and quantitatively analyzed using electronic nose combined with chemometric methods. Goat milk adulterated with proteins was detected by electronic nose and then the electronic nose response value was qualitative analyzed by principal component analysis (PCA), linear discriminant analysis (LDA), and the electronic nose response value was quantitative analyzed by linear regression analysis, Fisher discriminant analysis (FDA) and K nearest neighbor analysis (KNN). The results showed that: the principal component analysis and linear discriminant analysis are able to distinguish between different adulterations. The decision coefficient of linear regression analysis was 84.5%, which indicated that the reliability of the regression equation was high. The accuracy of the original classification of Fisher discriminant analysis reached 100%, and cross validation was 98.2%, indicating that the forecasting result is well. The classification accuracy of the training set of K nearest neighbor analysis is 95.1%, and the classification accuracy rate of the validation set is 97.1%, indicating that the prediction results of the model are well. All of these showed that electronic nose technology has certain feasibility in recognition of protein adulterations in goat milk.

Key words: goat milk, protein adulterations, electronic nose, chemometric methods