Chemiometry in the classification of chilled raw milk

Abstract

The objective was to evaluate the correlations between the quality parameters of chilled raw milk from different producing units through the principal components analisys (PCA). The following variables were considered: acidity, cryoscopy, density, fat content, total dry extract (TDE), defatted dry extract (DDE), pH, somatic cell count (SCC), total bacteria count (TBC) and  methylene blue reduction test (MBRT). It was observed that the samples were grouped into five different groups: (I) had no influence from any vector, (II) correlated with the SCC and TBC parameters, possibly because they have higher counts, signaling that they have a lower quality standard compared to the other samples, group (III) was more influenced by the  DDE, acidity, density and MBR parameters, while group (IV) was mainly discriminated by the TDE and fat vectors, indicating that the samples exhibit a higher content of these parameters and are produced from good quality milk, group (V) was characterized by a negative correlation with the vectors of TDE, fat, cryoscopy and pH, showing that these samples have a lower quality standard. From the PCA analysis, the samples were grouped according to similarity, showing that the vectors that most contributed to the discrimination of different milk samples were the SCC and TBC values.