摘要
本文介绍了一种新的多元回归分析方法——主成分回归分析。该法在自变量间存在多重共线性关系时,对原资料的主成分作回归分析,以消除回归系数中可能出现的不合理符号。并举例说明。
A new multiple linear regression method, principal component regression analysis, is introduced in this paper. The method can eliminate unresonable sign by constructing the regression of Y on principal components of independent variables X_i, while the relationship of X_i is multicollinearity. The method is illustrated with example.
出处
《中国卫生统计》
CSCD
北大核心
1991年第1期20-22,共3页
Chinese Journal of Health Statistics
关键词
回归分析
主成分分析
Principal component regression Regression analysis Principal components analysis Multicollinearity
