Molina E, Ramos M, Martin-Alvarez P J
Instituto de Fermentaciones Industriales, C.S.I.C., Madrid, Spain.
Z Lebensm Unters Forsch. 1995 Oct;201(4):331-5. doi: 10.1007/BF01192727.
Stepwise multiple linear regression (SMLR) and principal components regression (PCR) have been used to predict the percentages of cows', goats' and ewes' milk in ¿Iberico¿ cheese, using the results obtained by electrophoretic analysis (PAGE and IEF) of whey proteins, using standard cheeses. Similar predictions of the percentages of milks from the three species were obtained when either SMLR or PCR were applied to the electrophoretic data, i.e., the optical intensity of the electrophoretic bands (PAGE or IEF) of the whey proteins. The root mean square error of prediction in cross-validation (RMSEPCV) was lower than 4% in all cases.
逐步多元线性回归（SMLR）和主成分回归（PCR）已被用于预测伊比利亚奶酪中牛奶、羊奶和母羊奶的百分比，使用的是通过对乳清蛋白进行电泳分析（PAGE和IEF）并使用标准奶酪所获得的结果。当将SMLR或PCR应用于电泳数据（即乳清蛋白的电泳条带（PAGE或IEF）的光密度）时，得到了三种物种牛奶百分比的相似预测结果。在所有情况下，交叉验证中的预测均方根误差（RMSEPCV）均低于4%。
