论文部分内容阅读
人工神经网络具有对非线性系统预测的重要特性 ,其跟踪性能好 ,适用面广 ,收敛速度快 ,容错能力强。本研究利用陕西汉中地区 1 974～ 1 997年的病情、菌量、品种和气象资料 ,采用逐步回归法选择了影响汉中小麦条锈病流行的主要因子 ,即春季菌量、秋季菌量、感病品种面积比例、4月份降雨量和 4月份平均温度 ,并将其作为 BP神经网络预测模型的输入 ,用 1 974～ 1 993年的资料进行网络训练 ,对 1 994～ 1 997年小麦条锈病的流行程度作短期预测 ,结果高度吻合
Artificial neural network has the important characteristic of predicting nonlinear system, its tracking performance is good, wide range of application, fast convergence, fault-tolerant ability. In this study, the major factors influencing the epidemic of wheat stripe rust in Hanzhong were selected by using the disease, bacteria amount, variety and meteorological data from 1974 to 1997 in Hanzhong area of Shaanxi Province. The area proportion of varieties, the rainfall in April and the average temperature in April were used as the input of BP neural network prediction model. The data of 1 744-1993 were used for network training. The data of wheat stripe rust from 1994 to 1997 The prevalence of short-term forecasts, the results are highly consistent