ABSTRACT
Artificial neural network (ANN), an adaptive learning system, is able to learn patterns to develop a model from existing observations. ANN has been successfully applied to food processing systems including drying, baking, osmotic dehydration, high pressure processing, as well as estimations of a number of food properties and quality indicators. This chapter describes the principles behind ANN and the current status of its applications in food processing. ANN is an information processing system imitating biological neural networks in the brain. Many types of ANN have been developed, including multilayer perceptron network, Hopfield network and Kohonen network. Similar to multilayer perceptron network, a radial basis function (RBF) network is composed of three layers: input, hidden and output layers. However, transfer function between the hidden layer and the input layer is a RBF. ANN can also be regarded as a technique to achieve a particular learning task.
