摘要
本文针对一个因变量多元回归在气象预报应用中的不足和局限,引入多因变量多元回归,讨论了应用中的公共预报因子问题和因子筛选问题,并对多因变量回归与单因变量回归、场展开预报、逐步回归、逐段回归等方法进行了比较。
To over come the defect and the limit that were caused by single dependent variable regression model, this paper uses the multiple dependent variable matrix regression method to solve the problem of common predictors and the problem of screening predictors . The multiple dependent variable matrix regression forecast method is compared with other statistical regression methods. The characteristic and the effect using this method are also analyzed.
出处
《大气科学》
CSCD
北大核心
1997年第4期493-498,共6页
Chinese Journal of Atmospheric Sciences
关键词
多因变量
矩阵回归
统计预报
天气预报
multiple dependent variable common prediction factors matrix regression