Development of ANN-based sorption isotherm algorithm for the prediction of EMC of the paddy
A sorption isotherm algorithm based on artificial neural network (ANN) was developed for the prediction of equilibrium moisture content (EMC) of paddy under low temperature conditions (20 to 400C). ANN architecture was modeled by considering water activity (aw) and temperature (T) as input and EMC as output neurons. A sorption isotherm experiment under low temperature conditions viz. 20, 25, 30 and 35oC was conducted for providing training, testing and validation data of ANN. 2-7-1 was selected as the best ANN architecture on the basis of which sorption isotherm algorithm was developed in MATLAB R2015a. Mathematical modeling was also performed for the analysis of sorption isotherm behavior. Among four models as applied for the analysis, modified Chung-Pfost model showed best results with coefficient of determination (R2) value of 0.98 and mean square error (MSE) value of 8.82X10-05. Hence, this study involves a pioneering approach in the post harvest modeling aspect of sorption isotherm study of paddy.