摘要
对于正态线性模型的一般线性假设,讨论了实际中经常遇到的具有边界条件的线性假设的检验问题,利用几何投影工具得到了检验统计量。该方法比一般的代数方法要简单。
Regarding the general linear hypothesis of normal linear model, the problem of testing linear hypothesis with boundary conditions is discussed, which is often encountered in practice. The tool of geometric projection is applied to obtain the testing statistic. This method is much simpler than the general algebra methods.
出处
《南开大学学报（自然科学版）》
CAS
CSCD
北大核心
2009年第3期86-90,共5页
Acta Scientiarum Naturalium Universitatis Nankaiensis
基金
国家自然科学基金(10671099)
高等学校博士学科点专项科研基金(20050055038)
关键词
约束最小二乘解
边界条件
似然比
正交投影
矩阵广义逆
restricted least squares solution
boundary conditions
likelihood ratio
orthogonal projeetion
generalize inverse of matrices
