文献类型：期刊文献

中文题名：基于GA-Elman神经网络的参考作物需水量预测

英文题名：Prediction of Water Demand for Reference Crops Based on GA-Elman Neural Network Model

作者：李志新 赖志琴 龙云墨

第一作者：李志新

机构：[1]贵州理工学院土木工程学院

第一机构：贵州理工学院土木工程学院

年份：2019

卷号：0

期号：2

起止页码：117-120

中文期刊名：节水灌溉

外文期刊名：Water Saving Irrigation

收录：CSTPCD;;北大核心:【北大核心2017】；

基金：贵州省科学技术基金计划项目(黔科合基础[2016]1062);国家自然科学基金项目(51508121);贵州省科技合作计划项目(黔科合LH字2016[7096]);贵州理工学院大创项目(201814440070)

语种：中文

中文关键词：神经网络;GA-Elman;作物需水量;预测

外文关键词：Neural network;GA-Elman;crop water demand;prediction

摘要：根据通东灌区多年气象资料以及逐日参考作物需水量数据,构建了以日序数、日照时数、日平均气温等因子为输入向量,以逐日参考作物需水量为输出向量的GA-Elman神经网络参考作物需水量预测模型,通过随机选取方式,采用通东灌区2015年逐日气象资料及参考作物需水量等数据对模型进行了测试,结果表明:构建的GA-Elman模型具有较高的预测性能精度,其相对误差绝对值均值为7.24%,绝大多数处于0%～25%范围内,其中处于0%～10%范围内的占总数81.8%,模型具有较强的实际应用价值。
Based on the many years of meteorological data and daily reference crop water demand data in Tong Dong irrigated area,the GA-Elman neural network reference crop water demand forecasting model,with daily number,sunshine time and daily mean temperature as input vectors and daily reference crop water demand as the output vector,was constructed. The model was tested with the data of day by day meteorological data and the water requirement of reference crop of 2015 in Tong Dong irrigation area. The results show that the GA-Elman model has high predictive precision,the mean relative error of absolute value is 7.24%,the overwhelming majority is in the 0% ~ 25% range,and the total number in the 0 ~ 10% range is 81.8%. It has a strong practical application value.