基于学习参数梯度更新模型的农田水利灌溉管道灌水量预测方法  

A method for predicting irrigation volume of agricultural irrigation pipeline based on learning parameter gradient updating model

作　　者：周爱民 周钰 ZHOU Aimin;ZHOU Yu(Juye County Agriculture and Rural Bureau,Heze 274000,China;Juye County Natural Resources and Planning Bureau,Heze 274000,China)

机构地区：[1]巨野县农业农村局,菏泽274000 [2]巨野县自然资源和规划局,菏泽274000

出　　处：《上海农业学报》2025年第1期118-123,共6页Acta Agriculturae Shanghai

摘　　要：常规的农田水利灌溉管道灌水量预测方法,参数缺少有效更新,导致预测误差较大。因此,提出基于学习参数梯度更新模型的农田水利灌溉管道灌水量预测方法。使用SVM算法,计算函数估计的最小二乘值,并按照最优超平面参数,对数据进行预处理并更新,导入预处理后的农田水利灌溉管道灌水数据,归一化处理后对核函数进行确定。基于学习参数梯度更新模型建立模型二层结构,利用学习参数梯度构建灌水量预测模型,设置模型神经元个数,对数据进行训练,并计算得到农田水利灌溉管道灌水量数据预测结果。通过试验验证了本研究设计的农田水利灌溉管道灌水量预测方法,其预测误差较小,具有较好的应用价值。The conventional method for predicting the irrigation volume of agricultural irrigation pipelines lacks effective updates of parameters,resulting in significant prediction errors.Therefore,a method for predicting the irrigation volume of agricultural irrigation pipelines based on a learning parameter gradient updating model was proposed.Using SVM algorithm,the least squares value of the function estimation was calculated,and the data were preprocessed and updated according to the optimal hyperplane parameters.The preprocessed irrigation volume data of agricultural irrigation pipeline were imported,and the kernel function was determined according to the data after normalization.The two-layer structure of the model was established based on the learning parameter gradient updating model.Using the learning parameter gradient,an irrigation volume prediction model was constructed.The number of neurons in the model was set,and the data were trained and calculated to obtain the prediction results of irrigation data of agricultural irrigation pipeline.Further experiments were conducted to verify that the irrigation volume prediction method for agricultural irrigation pipeline designed in this study had a small prediction error and had good application value.

关 键 词：学习参数梯度更新模型 SVM算法 管道灌水量预测 农田水利灌溉 

分 类 号：S126[农业科学—农业基础科学] S607