回归模型中异方差或变离差检验问题综述

基金项目:

国家社会科学基金(02BTJ001).

详细信息

Developments of the Tests for Heteroscedasticity or Varying Dispersion in Regression Models

摘要: 回归模型的异方差或变离差检验是统计诊断的重要课题。本文系统介绍了普通回归模型、广义回归模型和基于纵向数据的随机效应或自相关回归模型的异方差检验或变离差检验的研究概况和最新进展;同时介绍了作者关于非线性回归模型的相应工作,最后指出了若干有有待进一步研究的问题。
- 非线性回归 /
- 广义非线性模型 /
- 指数族分布 /
- 异方差 /
- 变离差 /
- 偏大离差 /
- 随机效应 /
- 纵向数据 /
- 似然比检验 /
- score检验 /
- 自相关性 /
- 统计诊断
Abstract: The tests for heteroscedasticity or varying dispersion in regewssion models are very important problems of statistical diagnostics. This paper systematically surveys the studying developments of the tests for heteroscedasticity or varying dispersion in general regression models, generalized regression models and regression models with random effects based on longitudinal data. The work done by us in recent years are introduced and some problems about heteroscedasticity or varying dispersion are introducted. Several problems to be resolved are suggested.