Determination of the Betacyanin and Betaxanthin Contents of Red Beet (Beta Vulgaris) Powder Using Partial Least Square Regression Based on Visible-Near Infrared Spectra

Red beet (Beta vulgaris) contains betalain, which comprises red-violet betacyanin and yellow betaxanthin with esthetic and health benefits. Betacyanin and betaxanthin are usually detected using common chemical analysis, which requires a long time, skilled analysts, and sample destruction. For fast and accurate measurement, this study utilized a portable low-cost Visible-Near Infrared Spectra (Vis-NIR) spectrometer at 350-1000 nm combined with partial least square regression to predict the betacyanin and betaxanthin contents of red beet powder. The best calibration models for betacyanin and betaxanthin had R2c of 0.89 and 0.919, respectively, and standard error of calibration SEC of 0.108 and 0.037 mg/g, respectively. The models were able to predict the contents of both pigments with R2p of 0.87, standard error of prediction SEP of 0.108 mg/g, and the ratio of prediction to deviation RPD of 2.52 for betacyanin and R2p of 0.84, SEP of 0.056 mg/g, and RPD of 2.47 for betaxanthin. When applied to external unknown data, the models predicted the contents of betacyanin and betaxanthin with R2 of 0.98 and root mean square error of 0.107 and 0.055 mg/g. Moreover, the predicted values were not significantly different at 95 confidence. © 2024, Walailak University. All rights reserved.