正态广义非线性模型参数置信域的曲率表示

Confidence Regions in Terms of Curvature in Normal Generalized Nonlinear Model

本文从几何观点研究正态广义非线性模型参数和子集参数的置信域的曲率表示,由于本文研究的非线性模型中随机误差的协方差阵可以是正定和非负定的,因此,有关结论推广与发展了现有文献中的许多结论。

This paper studies approximate confidence regions for the parameter and parametersubsets in normal generalized nonlinear model from geometric point of view. Since covariancematrix of random error in the model is positive definite or un-negative definite,the resultsof this paper improves and develops many results in the literatures.