逐步回归分析法的一点不足之处  

Discussion on shortcomings of Stepwise Regression Analysis

作　　者：周文芳[1] 李民[1] 

机构地区：[1]武汉大学水利水电学院,武汉430072

出　　处：《西北水电》2004年第4期49-50,共2页Northwest Hydropower

摘　　要：在大坝安全监测分析中,通常采用统计模型、确定性模型或者混合模型对影响观测值的环境量进行定性和 定量分析而获得这些模型的最优解,一般通过逐步回归分析法。但事实上,在遇到具体情况时逐步回归分析法并不 一定能寻求到最优解,通过全面比较各因素对观测值的影响,提出了逐步回归分析法的一点缺陷。The statistic model, deterministic model and mixing model are usually proposed to qualitatively and quantitatively analyze the influence on environmental factors of observed value in the monitoring of dam safety. Moreover, the stepwise regression is considered to be able to achieve the optimum solution of these models. But in fact, the stepwise regression analysis may not be able to gain the solution value in some specific cases. This article presents shortcomings of stepwise regression analysis by overall comparison.

关 键 词：大坝安全监测 逐步回归分析 最优回归方程 

分 类 号：TV698.1[水利工程—水利水电工程]