Rapid discrimination of adulteration in Radix Astragali combining diffuse reflectance mid-infrared Fourier transform spectroscopy with chemometrics.

Fraud in the global food and related products supply chain is becoming increasingly common due to the huge profits associated with this type of criminal activity and yet strategies to detect fraudulent adulteration are still far from robust. Herbal medicines such as Radix Astragali suffer adulteration by the addition of less expensive materials with the objective to increase yield and consequently the profit margin. In this paper, diffuse reflectance mid-infrared Fourier transform spectroscopy (DRIFTS) was used to detect the presence of Jin Quegen in Radix Astragali. 900 fake samples of Radix Astragali produced by 6 different regions were constructed at the levels of 2%, 5%, 10%, 30% and 50% (w/w). DRIFTS data were analyzed using unsupervised classification method such as principal component analysis (PCA), and supervised classification method such as linear discrimination analysis (LDA), K-nearest neighbor (K-NN), linear discrimination analysis combining K-nearest neighbor (LDA-KNN) and partial least squares discriminant analysis (PLS-DA). The results of PCA showed that it was feasible to detect the adulteration of Radix Astragali by the combination of drift technique and chemometrics. PLS-DA obtained the best classification results in all four supervised methods with mean-centralization as the data preprocessing method, the prediction accuracy of PLS-DA model for the six groups of sample ranged from 95.00% to 98.33%. At the same time, LDA-KNN also achieved good classification results, and its correct prediction rate were also between 86.67% and 100.0%. The prediction results confirmed that the combination of DRIFTS technology and chemometrics can distinguish the amount of adulteration present in Radix Astragali. Additionally, the innovative strategy designed can be used to test the fraud of various forms of herbal medicine in other products.