基于神经网络的农业病虫害损失预测  

Prediction of Loss Affected by Pests Based on Neural Network

作　　者：刘璇[1,2] 唐慧强[1] 许遐祯[2] 项瑛[2] 

机构地区：[1]南京信息工程大学信息与控制学院,南京210044 [2]江苏省气象局,南京210044

出　　处：《农机化研究》2009年第4期13-16,共4页Journal of Agricultural Mechanization Research

基　　金：江苏省高校高新技术产业发展项目(JHB05-17)

摘　　要：鉴于农业病虫害经济损失的预测具有较强的复杂性和非线性特性,设计了一种新型的GRNN预测模型,对农业病虫害经济损失进行预测。该模型基于人工神经网络捕捉非线性变化独特的优越性,在神经网络技术和江苏省气象局提供的数据的基础上,利用MATLAB人工神经网络工具箱及GRNN广义回归神经网络建立预测模型,来提高农业病虫害经济损失预测的精度。预测结果表明,该方法建立的模型可以实现对病虫害经济损失的预测,且其预测精度较高。A new design of model for prediction of loss affected by pests has been accomplished by GRNN neural network aimed to solve the strong quality of complexity and nonlinear prediction . GRNN neural network is applied to predict economic loss because of its good quality of nonlinear prediction. And for the purpose of implementing the prediction accuracy, the factors affecting economic loss are analyzed by using the MATLAB neural network box to establish GRNN network based on neural network technology and the datum given by Jiangsu Meteorological Bureau. The model is istablished with this method, and the prediction of economic loss affected by pets is fulfilled;its prediction precision is very effective.

关 键 词：病虫害经济损失预测 GRNN神经网络 MATLAB BP神经网络 

分 类 号：S431[农业科学—农业昆虫与害虫防治] TP389[农业科学—植物保护]