Neural network classification of tan spot and Stagonospora blotch infection periods in a wheat field environment

Tan spot and Stagonospora blotch of hard red spring wheat served as a model system for evaluating disease forecasts by artificial neural networks. Pat hogen infection periods on susceptible wheat plants were measured in the fi eld from 1993 to 1998, and incidence data were merged with 24-h summaries o f accumulated growing degree days, temperature, relative humidity, precipit ation, and leaf wetness duration. The resulting data set of 202 discrete pe riods was randomly assigned to 10 model-development or -validation (n = 50) data sets. Backpropagation neural networks, general regression neural netw orks, logistic regression, and parametric and nonparametric methods of disc riminant analysis were chosen for comparison. Mean validation classificatio n of tan spot incidence was between 71% for logistic regression and 76% for backpropagation models. No significant difference was found between method s of modeling tan spot infection periods. Mean validation prediction accura cy of Stagonospora blotch incidence was 86 and 81% for backpropagation and logistic regression, respectively. Prediction accuracies of other modeling methods were less than or equal to 78% and were significantly different (P = 0.01) from backpropagation, but not logistic regression, results. The bes t back-propagation models of tan spot and Stagonospora blotch incidences co rrectly classified 82 and 84% of validation cases, respectively. High class ification accuracy and consistently good performance demonstrate the applic ability of neural network technology to plant disease forecasting.