单 位：
西北农林科技大学旱区农业水土工程教育部重点实验室
关键词：
参考蒸发蒸腾量;主成分分析;多元线性回归;BP神经网络
摘 要：
由于参考作物蒸发蒸腾量的影响因子众多,而且存在信息重叠,首先利用DPS软件对众多因子进行主成分分析,然后建立多元线性回归模型和BP神经网络模型对灌区参考作物蒸发蒸腾量进行预测,并将预测结果与Pen-man-Monteith公式计算值进行比较,结果发现多元线性回归模型预测的平均相对误差为10.05%,而BP神经网络模型预测的平均相对误差仅为2.71%,通过比较,BP神经网络模型能更好地满足灌区参考蒸发蒸腾量的预测要求。
译 名：
Research on Reference Crop Evapotranspiraion Prediction model in irrigation District based on Principal Component Analysis
作 者：
WEI Xin-guang,WANG Mi-xia,ZHANG qian(Key Laboratory of Agricultural Soil and Water Engineering in Arid Area of Ministry of Education,Northwest A﹠F University,Yangling712100,Shaanxi Province,China)
关键词：
reference evapotranspiration;principal component analysis(PCA);multiple linear regression;BP neural network
摘 要：
There are many factors that can influence the evapotranspiration of the reference crop,and there is also information overlap.In this paper,firstly,the principle component of many factors is analyzed by using the software of DPS,then the multiple linear regression and BP neural network model are established to predict the evapotranspiration of the reference crop,and the results is compared with the calculated values of Penman-Monteith equation.It is found that the average relative error of the multiple linear regression model prediction is 10.05%,while the average relative error of the BP neural network prediction model is only 2.71%.Through comparing the two models,it is found that the BP neural network model can better meet the prediction of the reference evapotranspiration irrigation requirements.
