Predicting aflatoxin contamination in peanuts: A genetic algorithm/neural network approach

Aflatoxin contamination in peanut crops is a problem of significant health and financial importance. Predicting aflatoxin levels prior to crop harvest is useful for minimizing the impact of a contaminated crop and is the goal of our research. Backpropagation neural networks have been used to model p roblems of this type, however development of networks poses the complex pro blem of setting values for architectural features and backpropagation param eters. Genetic algorithms have been used in other studies to determine para meters for backpropagation neural networks. This paper describes the develo pment of a genetic algorithm/backpropagation neural network hybrid (GA/BPN) in which a genetic algorithm is used to find architectures and backpropaga tion parameter values simultaneously for a backpropagation neural network t hat predicts aflatoxin contamination levels in peanuts based on environment al data. Learning rate, momentum, and number of hidden nodes are the parame ters that are set by the genetic algorithm. A three-layer feed-forward netw ork with logistic activation functions is used. Inputs to the network are s oil temperature, drought duration, crop age, and accumulated heat units. Th e project showed that the GA/BPN approach automatically finds highly fit pa rameter sets for backpropagation neural networks for the aflatoxin problem.