Artificial neural networks as a tool to predict the migration of silver from Antimicrobial nano-coated ldpe films

Antimicrobial nanocomposite silver (Ag) coated LDPE films were developed using a number of surface treatments to enhance attachment of Ag nanoparticles (NPs) and reduce potential migration of NPs into food. The migration of NPs into food has caused concern regarding the safety of NP food packaging, mainly surrounding human exposure and potential toxicity. Due to the complexity of the NP migration mechanisms, artificial neural networks (ANNs) were investigated as an alternative methodology to conventional mathematical models for migration prediction. Two ANNs were investigated; generalised regression neural network (GRNN) and multilayer perceptron neural network (MLPNN). The migration of total Ag from an experimental AgNP coated LDPE food packaging material following incubation for two storage conditions was found to range from 1.1 - 72 μg/dm2. Twenty percent of the dataset from the migration experiments was excluded from the ANN training set to allow for the model to be validated. Both ANNs were found to predict migration with R2 values ranging from 0.76 - 0.91 (GRNN) and 0.76 - 0.94 (MLPNN).