摘要
利用偏最小二乘回归方法对我国的粮食产量建立模型.在此基础上通过构造符合F分布的统计量,绘制T2椭圆图检测样本点中是否存在的异常点;通过构造变量投影指标反映每个自变量对因变量的影响程度,以此来提出提高因变量产量.
Author used partial least squares regression to make a model of grain output. Based on the model, author make a statistic which consistent with the distribution F. Drawing elliptical figure to detect the abnormal points from the sample points. Constructing variable projection index to reflects the degree of influence for the dependent variable.
出处
《韶关学院学报》
2011年第8期21-23,共3页
Journal of Shaoguan University
关键词
偏最小二乘回归
统计量
变量投影指标
模型分析
partial least squares regression
statistic
variable projection index
analysis of the model
