期刊文献+

提高回归模型拟合优度的策略(Ⅰ)——哑变量变换与其他变量变换 被引量:6

Strategy of improving the goodness of fit of the regression model (Ⅰ)——the transformation of the dummy variable and the other variable transformations
下载PDF
导出
摘要 本文目的是介绍第一种提高回归模型拟合优度的策略,即哑变量变换与其他变量变换。具体方法包括以下几个方面:①对多值名义自变量采取"哑变量变换";②对定量和有序自变量引入派生变量,包括"对数变换""平方根变换""指数变换""平方变换""立方变换"和"交叉乘积变换"的结果;③对定量因变量分别采取"对数变换""平方根变换""指数变换""倒数变换"和"Logistic变换";④构建回归模型时,在假定"包含截距项"与"不含截距项"的条件下,分别采取"前进法""后退法"和"逐步法"筛选自变量。得到了如下几个结论:①对定量因变量和自变量不做变量变换时,回归模型的拟合优度非常差;②根据资料所具备的条件,对定量因变量采取不同的变量变换方法,其回归模型的拟合优度是不尽相同的;③对多值名义自变量进行"哑变量变换"是常规的做法,但存在不足之处;④对定量自变量引入派生变量是非常有价值的;⑤假定回归模型中不含截距项有助于提高回归模型的拟合优度。 The purpose of this paper was to introduce the first strategy of improving the goodness of fit of the regression models,the transformation of the dummy variable and the other variable transformations.The concrete approaches were as follows:①“The transformation of the dummy variable”was adopted to the multi-value nominal independent variable.②The derived variables were introduced to the quantitative and the ordered independent variables including the results of“logarithmic transformation”“square root transformation”“exponential transformation”“square transformation”“cubic transformation”and“cross product terms transformation”.③“Logarithmic transformation”“square root transformation”“exponential transformation”“reciprocal transformation”and“Logistic transformation”were adopted to the quantitative dependent variable,respectively.④During building the regression models,the“forward selection”“backward selection”and“stepwise selection”were used for screening the independent variables under the conditions both with the intercept term and without it,respectively.The several conclusions were achieved as below:①The goodness of fit of the regression models was very poor when no transformations were applied to the quantitative dependent variable and independent variables.②The distinct results of the goodness of fit of the regression models could be achieved by using the distinct transformations to the quantitative dependent variable in accordance with the data conditions.③It was the common measurement to transform the multi-value nominal independent variable into the dummy variables,however,there were disadvantages of the approach mentioned above.④It was wonderful to introduce the derived variables to the quantitative independent variables in fitting the regression models.⑤It was helpful to improve the goodness of fit of the regression models by getting rid of the intercept term.
作者 胡良平 Hu Liangping(Graduate School,Academy of Military Medical Sciences PLA China,Beijing 100850,China;Specialty Committee of Clinical Scientific Research Statistics of World Federation of Chinese Medicine Societies,Beijing 100029,China)
出处 《四川精神卫生》 2019年第1期1-8,共8页 Sichuan Mental Health
基金 国家高技术研究发展计划课题资助(2015AA020102)
关键词 变量变换 哑变量变换 Logistic变换 派生变量 拟合优度 Variable transformation Transformation of the dummy variable Logistic transformation Derived variable Goodness of fit
  • 相关文献

参考文献1

二级参考文献4

共引文献10

同被引文献21

引证文献6

二级引证文献5

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部