Abstract:

Abstract: This article in view of the rice blast to rice growth process, by using the rice blast information of Zizhong, Sichuan province from 1998 to 2008, a rice blast prediction model is built based on gray artificial neural network. The model’s result shows that: the prediction accuracy of gray artificial neural network model is 0.0946, which is better than that of GM(1,1) model which is 1.8857. Gray artificial neural network model is feasible to fit all kinds of functional relationship, what’s more, it can make full use of the information and avoid the information distortion caused by positive and negative cancellation in the process of accumulative sequence of system identification. Gray artificial neural network model has good effects on fitting and forecasting precision, so the gray artificial neural network model can be used in the forecasting work of the occurrence of rice blast.