Abstract: In this study statistical methodology is used to determined similar varieties of Amaranthus varieties and to model the behaviour of the moisture content as a function of the water activity (a ). Different equationsw proposed in the literature are studied. When analyzing the equations relating the moisture content (m.c.) and the a , estimates of the parameters were obtained by fitting a nonlinear model for each variety Using the criteria w of minimum S, E or P we conclude that the G. A. B. (Guggenheim andersen and de Boer) and Oswin equation provide the best fits. For Amaranthus Cruentus the best fit is given by Oswin and Bradley model, according to P and E criteria, respectively. However, the order of the difference with respect to the goodness of fit statistics using G. A. B. model is negligible. Since the S goodness of fit criterion seems to be reasonable to select the equation in this setting, we use the S measure in order to choose a model. The S statistic for the G. A. B. model is 60% larger on the for Amaranthus Cruentus than on the Amaranthus Mantegazzianus.
A. M. Bianco , G. Boente , M. L. Pollio and S. L. Resnik , 2005. Sorption Isotherms of Two Varieties of Amaranthus at 25?c: Comparison of Water Sorption Characteristics and Mathematical Models . Journal of Food Technology, 3: 294-299.