In recent years, the combination of genistein (GEN) and curcumin (CUR) has attracted much attention due to its potential activity as an immunomodulatory agent. However, it is important for ensuring the quality and safety of the products using a simple, rapid, and cost-effective analytical method. UV-Vis spectroscopy method in combination with chemometrics techniques was successfully exploited to analyze the combination of GEN and CUR. Multivariate calibration models, namely, principal component regression (PCR) and partial least squares regression, were generated in this study using the R statistical software. It was found that the best predictive models for the quantitative determination of GEN and CUR were PCR on Savitzky-Golay smoothing spectral and PCR of original spectra, respectively. The sparse partial least square-discriminant analysis model was built with parameter tuning. The classification of GEN, CUR, their binary mixture, and methanol was successfully conducted by tuning the model parameters, including the prediction distance, number of components, and the keepX variables.
Key words: curcumin; genistein; multivariate calibration; discriminant analysis; spectroscopy
|
