【摘要】 Monlezun和Speed用正交投影的方法给出了线性约束下线性模型的参数估计和假设检验的统计量。本文仅将其对多元线性模型的情形作一推广。这里不仅避免了经典方法中的Lagrange乘子法,而且对假设检验的统计量的推导也比经典方法更简捷。更多还原

【Abstract】 C. J. Monlezun and F.M. Speed have given the geometric approach to estimation and hypotnesis testing in the constrained linear model byorthognally projecting. In this paper we want to generalize such a technique in multivariable linear model and there is no need to employ the language Multiplier technipue. The quadratic form for testing the null hypothesis is more computationally efficient than classical.更多还原