李长平,胡良平.非配对设计二值资料一水平多重Logistic回归分析[J].四川精神卫生杂志,2019,32(4):297-303.,One-level multiple Logistic regression analysis with the dichotomous choice data collected from the unpaired design[J].SICHUAN MENTAL HEALTH,2019,32(4):297-303
非配对设计二值资料一水平多重Logistic回归分析
One-level multiple Logistic regression analysis with the dichotomous choice data collected from the unpaired design
投稿时间:2019-08-01  
DOI:10.11886/j.issn.1007-3256.2019.04.002
中文关键词:  二值资料  一水平  派生变量  加权回归  多重Logistic回归分析
英文关键词:Binary data  One-level  Derived variables  Weighted regression  Multiple Logistic regression analysis
基金项目:国家高技术研究发展计划课题资助(2015AA020102)
作者单位
李长平 天津医科大学公共卫生学院卫生统计学教研室世界中医药学会联合会临床科研统计学专业委员会 
胡良平 世界中医药学会联合会临床科研统计学专业委员会军事科学院研究生院 
摘要点击次数:
全文下载次数:
中文摘要:
      【摘要】 本文的目的是介绍非配对设计二值资料一水平多重Logistic回归模型的构建与求解方法。基于SAS软件分别对以列联表和数据库形式呈现的定性资料进行全面分析,并得出了4个对提高模型拟合优度很有价值的结论:第一,若资料以列联表形式呈现,应拟合“加权”Logistic回归模型;第二,若资料中包含定量自变量,不适合将其定性化;第三,若资料中包含定量自变量,应依据定量自变量和二值自变量产生出派生自变量;第四,若资料中有定性自变量时,必须将多值名义或有序自变量进行哑变量变换,不需要依据二值自变量产生出派生自变量。
英文摘要:
      The purpose of this paper was to introduce the construction and solution of one-level multiple Logistic regression model with binary data collected from the unpaired design. Based on SAS software package, the qualitative data presented in the form of contingency tables and databases were analyzed comprehensively and thoroughly, and four valuable conclusions were obtained to improve the goodness of fit. First, if the data were presented as contingency tables, the weighted Logistic regression model should be fitted. Second, if the data contained quantitative independent variables, which were not suitable for transforming into qualitative variables. Third, if the data contained quantitative independent variables, derived independent variables should be generated according to quantitative independent variables and binary independent variables. Fourth, if there were qualitative independent variables in the data, the multi-valued nominal or ordered independent variables should be transformed into dummy variables, and the derived independent variables did not need to be generated according to the binary independent variables.
查看全文  查看/发表评论  下载PDF阅读器
关闭