论文部分内容阅读
通过对平凉市近25年小麦条锈病及其气象数据的分析,选择了主要的影响因子,采取改进BP神经网络算法对小麦条锈病菌在越夏期感染程度、秋季麦苗感染程度及翌年春季感染程度分别建立了预测模型进行了预测。仿真结果显示3个BP预测模型均达到0.000 1的训练目标,预测准确度达到100%,测试样本预测输出与实际输出间的均方差为0.1～0.4级。由此可见,改进BP网络预测模型比逐步回归方法具有更好的预测精度,能满足应用需求。
Based on the analysis of wheat stripe rust and its meteorological data in Pingliang City in the past 25 years, the main influencing factors were selected, and the improved BP neural network algorithm was used to analyze the infection degree of wheat stripe rust in summer, the degree of wheat seedling infection in autumn and the degree of infection in next spring Forecast models were established respectively. The simulation results show that all three BP prediction models reach the training target of 0.000 1, the prediction accuracy reaches 100%, and the mean square error between the predicted output and the actual output of the test sample is 0.1 to 0.4. Thus, it can be seen that the improved BP network prediction model has better prediction accuracy than the stepwise regression method and can meet the application requirements.