Efficient and rapid screening of camel Milk powder adulteration using Raman spectroscopy with a hierarchical discrimination strategy model.

This study developed a rapid and efficient method based on Raman spectroscopy and a hierarchical discrimination strategy for screening adulteration in camel milk powder. Raman spectra of pure camel milk powder and 3 common adulterants were collected and analyzed using chemometric methods to establish a "purity-type-concentration" hierarchical discrimination model. Firstly, a partial least squares-discriminant analysis (PLS-DA) model was used to accurately identify adulterated samples, achieving a prediction sensitivity of 1.000 and specificity of 0.987. Secondly, a multi-class classification model was established to identify the type of adulterant, with particularly high accuracy for soybean milk powder-adulterated samples (sensitivity 0.969, specificity 0.996). Finally, separate classification models for adulteration concentration levels were developed for each adulterant category, with discrimination performance varying considerably among adulterant types: good performance was obtained for low-level adulteration (5%/10%), whereas the sensitivity for certain concentration levels of goat milk powder-adulterated samples was as low as 0.111, under the experimental conditions investigated. Within the scope of the samples investigated, this method enables authenticity verification and adulterant type identification of camel milk powder, as well as adulteration level assessment for most adulterant types and concentration levels, with potential for application in food quality control and market supervision.