【摘要】 近年来,对于回归分析中复共线性问题的研究。一直占有相当重要的地位。许多统计工作者在这一领域内进行了深入的研究。并取得了非常丰富的成果。本文在他们研究的基础上,系统地阐述了复共线性产生的原因及消除这种复共线性所采取的种种方法,并且富有成效地引入了线性模型中回归系数估计的一种估计类,使得一些非常重要的估计,如最小二乘估计、岭估计、Stein估计等都在这一估计类中得到了统一的描述。最后,本文还通过理论分析丛实例演示论证了在这个估计类中,各种回归系数估计有着和其他工作者所得到的相容的结论。更多还原

【Abstract】 In recent years, the researchment of multiliearity accounts for a very important position in regression analysis. Many statisticians have done extensive and detailed study in this field, and have attained very rich results. On the basis of their research work, we explain systematically the resons of Multilinearity and the methods of eliminating Multi-linearity, and we also introduce effectively a kind of estimate class of the regression coefficient in linear model, many important estimates, for example, LSE, Ridge estimate, stein estimate etc. . have a uniform interpretation in this class. Finally, through the theory analysis and the example demonstration, we prove from this estimate class that we get the results which are interlinked with the results attained by other statisticians.更多还原