Rapid Detection and Quantification of Adulterants in Camel Milk Powder Using Near-Infrared (Nir) Spectroscopy with Chemometrics

Abstract

driven by economic interests, camel milk powder (CMP) adulteration become a high-frequency issue in China. In this study, NIR spectroscopy was coupled with chemometric methods to detect and quantify CMP and its adulterants including protein powder (PP), goat milk powder (GMP), and cow milk powder (MP) with different levels. Color analysis of the authentic and adulterated CMP samples was also performed. The results showed that the PCA with the first derivative NIR spectra showed better separation of different adulterations than the PCA with color data. The samples were classified with a sensitivity and specificity of 100% with the data-driven soft independent modeling of class analogy (DD-SIMCA) model. Compared with the LDA and KNN algorithms, the RF better discriminated the adulterated CMP samples with an accuracy of 98%. The PLSR models had the capability to accurately quantify the adulterated samples with the RMSEP ≤ 0.0481 and R2P ≥ 0.9767.

Keywords: Camel milk powder, NIR spectroscopy, Chemometrics, Adulteration, Detection, quantification