摘要
用马氏距离客观确定样本的空间结构后,我们可定义在E最小意义下高维空间到低维空间的映射。根据这个原理,L(L>2)个预报方法可综合为2个新因子,最后通过点聚图分析作出预报决断。
After the space structure is fixed objectively with Mahalanobis's distance the image from a higher-dimensional space to a lower-dimensional space is defined based upon the minimum E. According to this theory L forecasting methods (L>2) may be synthesized to 2 new factors and the final conclusion is obtained with identifying drawing.
出处
《气象科学》
CSCD
北大核心
1991年第2期214-218,共5页
Journal of the Meteorological Sciences
