摘要
本文从预报对象残差和预报因子残差之间的相关性出发,论证了备选因子集中的因子应该由两部分组成:一部分为与预报对象之间的相关性较好而因子之间的相关性较差或适中的因子,另一部分为与预报对象之间的相关性较差而与主因子之间的相关性很差或很好的因子;提出了选取备选因予集中因子的一般思路——配合主因子法。
Discussions of the correlation between the residuals of predictand and predictor suggested that the candidate predictors consist of two different parts of Predictors.Predictors of the first part are these that the correlation between the predictand and predictor is better, while the correlation between one and another of the predictors is not good.Predictors of the second part are these that the correlation between the predictapd and predictor is very bad while the correlation between the main predictor and the predictors is best or worst. And a method of selecting predictors which are made up of the candidate predictors is put forward and called the Method suited to the main predictor.
出处
《气象》
CSCD
北大核心
1990年第2期16-22,共7页
Meteorological Monthly
