基于主成分回归的茎直径动态变化预测方法  

Prediction of Stem Diameter Variations Based on Principal Component Regression

作　　者：员玉良[1,2] 盛文溢[1] 

机构地区：[1]中国农业大学信息与电气工程学院,北京100083 [2]青岛农业大学机电工程学院,青岛266109

出　　处：《农业机械学报》2015年第1期306-314,共9页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家科技部国际合作资助项目(2010DFA34670);中国农业大学研究生科研创新专项资助项目(2013YJ008);中德联合研究小组资助项目(GZ494)

摘　　要：影响植物茎直径变化的因素有很多,除了植物的自然生长外,气象因子和土壤含水率也是十分重要的因素。以空气温度、相对湿度、气压和光合有效辐射4个温室内的主要气象因子和土壤含水率为观测对象,对处于生长末期的4株温室向日葵样本和2株西红柿样本进行监测试验。以其中一株向日葵样本为对象,对其茎直径变化的影响因素作主成分分析并建立回归模型。将试验样本上的监测数据输入模型,对向日葵样本和西红柿样本的茎直径变化量进行预测,并分别与其各自实测值比较。结果显示,该回归模型对处于生长末期的温室向日葵和西红柿茎直径动态变化有较好的预测,预测值与实测值相关分析的决定系数为0.649-0.782,均方根误差为0.029-0.143。Among the various factors affecting the variation of plant stem diameter, meteorological conditions and soil water content are very important ones, besides natural growth. Soil water content together with four main meteorological parameters in greenhouse, including air temperature, relative humidity, pressure and photosynthetically active radiation, were selected for observing with four sunflower samples and two tomato samples at late stage of growth. Using part of the data measured from one sunflower sample, the principal component analysis was performed to set up a regression model. Data from sunflower samples and tomato samples were input to the model to predict the stem diameter variations of the sunflower samples and tomato samples and compared with the observed stem diameter variations. Comparison results showed that the regression model had a good prediction for the dynamics of stem diameter variations in sunflowers and tomatoes at late growth stage. The coefficients of determination in correlation analysis were above 0.6 and reached 0. 649 - 0. 782, while the root mean square errors were 0. 029 - 0. 143.

关 键 词：茎直径 动态变化 预测 主成分分析 回归 

分 类 号：S24[农业科学—农业电气化与自动化] S161.21[农业科学—农业工程]