| 散度偏大计数数据回归模型的变量选择与模型比较 |
| |
| 引用本文： | 陈雪东.散度偏大计数数据回归模型的变量选择与模型比较[J].应用数学与计算数学学报,2013(4):533-540. |
| |
| 作者姓名： | 陈雪东 |
| |
| 作者单位： | 湖州师范学院理学院,浙江湖州313000 |
| |
| 基金项目： | 国家社会科学基金资助项目（10BTJ001）;国家自然科学基金资助项目（11171105,11171293） |
| |
| 摘 要： | 讨论了具有散度偏大特征计数数据的建模与拟合问题.针对导致数据散度偏大的原因和常用的几类候选模型的结构,分别给出了关于嵌套模型的模型与变量同时选择的Bayes方法和关于非嵌套模型的模型检验与比较方法,并在此基础上进一步完善,提出了较为系统完整的模型与变量选择方法.实际例子说明了方法的具体实现过程和有效性.
|
| 关 键 词： | 散度偏大 嵌套模型 模型选择 Bayes因子 MCMC （Markov chain Monte Carlo）方法 |
Variable selection and models comparison of regression model on count data with overdispersion |
| |
| CHEN Xue-dong.Variable selection and models comparison of regression model on count data with overdispersion[J].Communication on Applied Mathematics and Computation,2013(4):533-540. |
| |
| Authors: | CHEN Xue-dong |
| |
| Institution: | CHEN Xue-dong (School of Science, Huzhou Teachers College, Huzhou 313000, Zhejiang Province, China) |
| |
| Abstract: | The problems of modeling and fitting on count data with overdisper- sion are disscussed. Accouting to the various reasons of count data exhibiting overdispersion and the different candidate models to fitting, a Bayesian procedure for the nested model and a test method for the non-nested model are proposed, respectively. A unified framework for the model comparison and the variable se- lection is provided. The example illustrates the implement and the effection of the method. |
| |
| Keywords: | overdispersion nested model model selection Bayes factor MCMC（Markov chain Monte Carlo） method |
| 本文献已被 维普 等数据库收录！ |
