| 相关二分类反应变量资料的Binomial-Poisson层次模型分析 |
| |
| 引用本文： | 曹文君,李运明,陈长生,李晨.相关二分类反应变量资料的Binomial-Poisson层次模型分析[J].中国卫生统计,2008,25(6). |
| |
| 作者姓名： | 曹文君 李运明 陈长生 李晨 |
| |
| 作者单位： | 1. 第四军医大学军事预防医学系卫生统计学教研室,710032
2. 四川大学华西公共卫生学院预防医学系2004级3班,610041 |
| |
| 基金项目： | 国家自然科学基金资助项目
，
陕西省自然科学基金资助项目
，
陕西省科技计划项目
|
| |
| 摘 要： | 目的探讨数据存在相关性时,二分类反应变量资料的统计分析方法。方法构建Binomial-Poisson层次模型,并结合毒理学实例与logistic回归模型进行比较。结果随机效应logistic回归模型和Binomial-Poisson层次模型都考虑了数据间的相关性,假设检验结论一致,但与传统logistic回归模型结论相反,而Binomial-poisson层次模型结果更加稳定。结论层次结构模型可有效分析相关二分类反应变量数据资料。
|
| 关 键 词： | logistic回归模型 Binomial-Poisson层次模型 二分类反应 过度离散 |
A Hierarchical Binomial-Poisson Model for the Analysis of Related Binary Reponses |
| |
| Cao Wenjun,Li Yunming,Chen Changsheng,et al..A Hierarchical Binomial-Poisson Model for the Analysis of Related Binary Reponses[J].Chinese Journal of Health Statistics,2008,25(6). |
| |
| Authors: | Cao Wenjun Li Yunming Chen Changsheng |
| |
| Abstract: | Objective To explore the statistical method in analysis of related binary responses.Methods Compared with logistic regression model,a Binomial-Poisson hierarchical model is developed to analyze the toxicological data.Results In logistic regression with random effects model and the Binomial-Poisson model we adjusted for the correlation among data,the results of hypothesis testing are comparable.The results of the Binomial-Poisson model are more stable.However,the common logistic regression model reaches an opposite conclusion.Conclusion Hierarchical models are effective in the situation involving grouped binary data in which the observations within individual groups are correlated. |
| |
| Keywords: | Logistic regression model Hierarchical Binomial-Poisson model Binary response Overdispersion |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
