论文部分内容阅读
【目的】建立基于小波神经网络病虫害预测预报模型,对提前采取防病防虫措施、减少农作物病虫害损失、提高农作物产量与质量具有重要意义。【方法】本研究以山西省运城市芮城县1980-2014年麦蚜发生程度和气象因子数据为基础,采用主成分分析法从40个基础气象因子中整合形成9个新的自变量输入模型,采用试凑法筛选隐含层节点数,用1980-2009年的数据进行网络训练,对2010-2014年麦蚜发生程度进行回测,建立了以Morlet小波函数为传递函数的小波神经网络模型,并与以Sigmoid函数为传递函数的BP神经网络模型进行了比较。【结果】小波和BP神经网络两种模型对训练样本的平均拟合精度均有10年以上超过80%,两者MAPE值分别为89.83%和83.07%,MSE值分别为0.0578和0.6192。【结论】两个模型都能较好地描述麦蚜发生程度;从预测精度和模型的稳定性来看,小波神经网络好于BP神经网络。
【Objective】 The objective of this paper is to establish a model of pest forecast based on wavelet neural network. It is of great significance to take measures to prevent diseases and pests in advance, reduce crop pests and losses, and improve crop yield and quality. 【Method】 Based on the data of wheat aphid and meteorological factors in Ruicheng County, Yuncheng City, Shanxi Province from 1980 to 2014, nine new input models of independent variables were integrated from 40 basic meteorological factors by principal component analysis , The trial-and-error method was used to screen the hidden layer nodes, and the data were used to train the network from 1980 to 2009. The wavelet neural network model , And compared with BP neural network model which uses Sigmoid function as transfer function. 【Result】 The average fitting precision of wavelet and BP neural network to training samples were more than 80% for more than 10 years. The MAPE of the two models were 89.83% and 83.07% respectively, and the MSE values were 0.0578 and 0.6192 respectively. 【Conclusion】 Both models can describe the occurrence of wheat aphid better. From the perspective of prediction accuracy and model stability, wavelet neural network is better than BP neural network.