The authors of a recent study (here – purchase required) investigated the effect of using different chemometric models to differentiate rice varieties using the same multivariate technique (Raman spectroscopy, also using FT-IR as a comparator) and the same reference sets of authentic rice. They found a significant difference between method performance depending on the classification model underpinning the chemometrics. They reported the optimum model to be Supervised Kohonen Map with Multiplicative Scatter Correction (SKM + MSC). They have proposed a Raman spectroscopy protocol which gives a reliable and rapid point-of-use screening test for rice grain variety.
Photo by Pierre Bamin on Unsplash