摘要
根据朝阳气象站的实测气象数据(温度、湿度、日照时数、风速、蒸发量、降水量),提出基于偏最小二乘回归的投影寻踪耦合模型用于预测朝阳地区参考作物滕发量。偏最小二乘回归方法能够有效地处理自变量间多重线性相关问题,但对处理因变量与自变量间复杂的非线性问题较差,而投影寻踪回归模型有效解决了非线性问题。把这2种方法结合在一起,建立了基于偏最小二乘回归的投影寻踪耦合模型,用于该地区参考作物滕发量的预测。并将耦合模型预测的ET0结果与Penman-Monteith公式计算的ET0结果进行比较,该耦合模型预测精度较高。
According to the meteorological data obtained by Chaoyang Weather Station（temperature,humidity,sunshine duration,wind speed,evaporation,precipitation） and the coupling model of projection pursuit based on partialeast-square regression used to predict the reference crop evaporation of Chaoyang City.The method of partial east-squares regression can effectively deal with the problem of multicollinearity among independent variable,but can not ideally solve the complicated problem of nonlinearity between dependent variables and independent variables,projection pursuit regression model can effectively deal with the problem of nonlinearity.This paper combines the two methods to establish the method of projection pursuit based on partialeast-square regression to predict the reference crop evaporation of the region.And the coupling models predict ET0 results with Penman-Monteith equation calculated ET0 results compares,the results show that the coupling models can predict wih a higher degree of accuracy.
出处
《中国农村水利水电》
北大核心
2011年第2期76-78,共3页
China Rural Water and Hydropower
关键词
偏最小二乘回归
投影寻踪
耦合模型
参考作物腾发量
腾发量预测
partialeast-squares regression
projection pursuit
coupling model
reference crop evaporation
evaporation predict
