Identification of ginseng according to geographical origin by near-infrared spectroscopy and pattern recognition

The feasibility of combining near-infrared (NIR) spectroscopy with chemometrics was explored to discriminate ginseng geographical origins. A total of 326 ginseng samples from three major ginseng producing regions were prepared and analyzed. After spectral pre-treatment, principal component analysis (PCA) was used for a preliminary analysis. Three algorithms, i.e., partial least squares-discriminant analysis (PLS-DA), soft independent modeling of class analogy classification (SIMCA) and successive projection algorithms-linear discriminant analysis (SPA-LDA), were applied to build models to discriminate origins of samples. The results showed that ginseng could be classified based on geographical origins with pattern recognition. By comparison, SPA-LDA is better than PLS-DA and SIMCA. It indicates that NIR spectroscopy combined with SPA-LDA is a potential and feasible tool for identifying ginseng according to geographical origin, but the effectiveness needs to be verified further.