摘要
作物需水量的预测是进行水资源规划和管理的有效手段。它与气象因子之间存在着严重的非线形关系。建立4个输入单元和1个输出单元的三层BP网络,选取不同的隐层结点数进行训练,并通过比较其相对误差的大小确定了神经网络的结构。利用MATLAB环境,提出基于BP神经网络的水稻需水量预报模型,并结合实际数据进行了检验。结果表明:该方法能够较好地反映气象因子与水稻需水量之间的关系,收敛速度快,预报精度较高。
Crop water requirement forcasting is the effective way to plan and manage the water resources. There is a serious nonlinear relationship between the crop water requirement and atmosphere factors. Three layers of BP networks with four importing units and one exporting unit were set, different hidden layer nodes were selected to train and then the relative errors were compared so as to confirm the neural network structure. The predictive model of water requirement of paddy base on MATLAB BP neural network was put forward in the paper, and the verification of the model was made by comparing the predictive value with the practical data. The results showed that the method could reflect the relation of crop water requirement and atmosphere factors. Moreover, the method had fast convergence speed and remarkable accuracy.
出处
《沈阳农业大学学报》
CAS
CSCD
北大核心
2005年第5期599-602,共4页
Journal of Shenyang Agricultural University
关键词
神经网络
水稻
需水量
预测模型
neural network
paddy
water requirement
predictive model
