Abstract
Wild honeys in Indonesia are still widely believed to be good for health with high economic value. This honey is naturally produced by Apisdorsata bee. In this study, authentication analysis by classification and discrimination of attenuated total reflectance-fourier infrared spectroscopy (ATR-FTIR) spectra was conducted on several wild honeys from various places in Indonesia (n = 186) which then compared to adulterated honey contained commercial sugars of aren (Arenga pinnata), coconut, and cane sugar at 10-50% concentration (n = 57). Combination of spectra measurement at 4,000-650 cm(-1) with Chemometric technique by several multivariate analyses resulted in visualization of honey grouping, classification, and regression model that differentiate these honeys, both partial and overall. Principle component analysis multivariate analysis was able to visualize the differentiation of adulterated honey from the authentic ones. Discriminant analysis, a supervised classification technique, was used to differentiate the fake from the authentic honey among those from various origins at wave number range of 4000-800 cm(-1) with performance index of 91,8, 90.32-100% sensitivity, and 95. 70-100% specificity. Partial least-squares analysis was used to build a model provided quantitative results of commercial sugars content in honey allegedly added during adulteration. Authentic honeys had commercial sugars content less than 10% with R(2) of aren, coconut, and cane sugar of 0.9995, 0.9980 and 0.9998, respectively, with their predictive R(2) values of 0.9977, 0.9983 and 0.9946, respectively.