Ulluco starch could be a promising renewable source for the production of biodegradable or edible films, as an alternative to plastic. This would mitigate the negative impact of plastics on the environment. The aim of this work was to study the effect of the starch concentration (SC), glycerol concentration (GC), and drying temperature (T) of ulluco starch-based films on their physical properties using stepwise regression (SR) and artificial neural network (ANN) approaches. The physical properties, such as the solubility (S), water vapour permeability (WVP), tensile strength (TS), elongation at break (EB), and transparency, of the edible films were satisfactory. The feed-forward and cascade-forward neural networks satisfactorily modelled the effect of the SC, GC, and T on the mechanical, optical, and water-affinity properties (WAP) of the edible films. ANN approach showed better results than SR in all the properties and ANN models were used in the sensitivity analysis and optimization. The global sensitivity analysis showed that the GC had the greatest influence on the physical properties. A desirability function-based optimization including WVP, EB and OP showed comparable values between experimental and estimated data. Based on the results of this study, the use of ulluco starch for the preparation of edible films has enormous potential for the replacement of non-biodegradable plastic packaging.
Keywords: Artificial neural network; Biodegradable films; Mechanical properties; Optical properties; Starch; Water affinity properties.
Copyright © 2018 Elsevier B.V. All rights reserved.