Aflatoxin Prediction Using a GA Trained Neural Network

Predictingthe level of aflatoxincontaminationin cropsof peanuts is a task of significant importance. Backpmpagationneural networkshavebeenused in the past to modelthis problem,but use of the backpropagationalgorithmfor training introduces limitations anddifficulties. Therefore,it is useful to explore alternative learning algorithms. Geneticalgorithmsprovidean effective techniquefor searching large spaces, and havebeen usedin the past to train neuralnetworks.Thispaperdescribesthe developmentof a genetic algorithm/neural networkhybrid in whicha geneticalgorithmis usedto find weightassignmentsfor a neural networkthat predicts aflatoxin contaminationlevels in peanutsbasedonenvironmentaldata.

• Type: Conference
• Year: 1998
• Where: FLAIRS
• Authors: C. E. Henderson, Walter D. Potter, Ronald W. McClendon, Gerrit Hoogenboom
• Keywords: Aflatoxincontaminationin Cropsof Peanuts, Alternative Learning Algorithms, Artificial Intelligence, FLAIRS 1998, Training Introduces Limitations