摘要
目的：建立一个较优的ｌｏｇｉｓｔｉｃ回归方程，探讨ｌｏ－ｇｉｓｔｉｃ回归方程中自变量的变换问题．方法：利用图示方法和Ｍｏｎｔｅ－ｃａｒｌｏ方法讨论了ｌｏｇｉｓｔｉｃ回归中自变量的变换问题．结果：对原有的图示方法进行了改进，提出了一种新的图示方法普通偏残差图，并用实例说明了其医学应用价值．结论：普通偏残差图不仅简单，而且能较准确地提示自变量的变换形式．
Aim:For fitting a good regression equation, the transformation of explanatory variables in logistic regression models was studied. Methods:The graphical methods and Monte carlo methods were used. Results:A new graphical methodgeneral partial residual plot was presented and demonstrated with a medical example. Conclusion:General partial residual plot is easy and effective in suggesting the form of transformation of explanatory variables.
出处
《第四军医大学学报》
1998年第3期254-256,共3页
Journal of the Fourth Military Medical University
基金
第四军医大学科研基金
关键词
变量变换
LOGISTIC回归
普通偏残差图
transformation of variables logistic regression general partial residual plot
