A novel method for shelf life prediction of a packaged moisture sensitive snack using multilayer perceptron neural network

Abstract

A novel method for shelf life prediction was established for a packaged moisture sensitive snack. Artificial neural network (ANN) based on multilayer perceptrons (MLP) with back propagation algorithm was developed to predict the shelf life of packaged rice snack stored at 30 °C and 75% RH, 30 °C and 85% RH and 40 °C and 75% RH, comparable to tropical storage conditions. The MLP predicted shelf lives were then compared to the actual shelf lives. Using MLP algorithm, many factors could be incorporated into the model including food characteristics, package properties, and storage environments. The MLP neural network comprised an input layer, one hidden layer and an output layer. The network was trained using Lavenberg–Marquardt (LM) algorithm. The performance of a MLP neural network was measured using regression coefficient (R2) and mean squared error (MSE). The MLP algorithm gave R2 of 0.98, and MSE of 0.12. MLP offers several advantages over conventional digital computations, including faster speed of information processing, learning ability, fault tolerance, and multi-output ability.