关于求解带约束秩亏线性回归方程组最小二乘解的一个算法  

AN ALGORITHM ON SOLVING LEAST SQUARES SOLUTION OF RANK-DEFECTIVE LINEAR REGRESSION EQUATIONS WITH CONSTRAINTS

作　　者：孙振东[1] 赵维谦[2] 

机构地区：[1]北京航空航天大学七研,北京100083 [2]青岛海洋大学应用数学系,青岛266003

出　　处：《高等学校计算数学学报》1995年第3期252-257,共6页Numerical Mathematics A Journal of Chinese Universities

摘　　要：<正> 1 引 言 考虑带线性约束的秩亏线性回归模型: Y=Xβ+ε, Hβ=c, (M_1) ε～N(0,σ~2V), V≥0,及带线性约束和非负性约束的秩亏线性回归模型:In this paper, we consider two kind of rank-defective linear regression models with constrains. The condition of definite solution which satisfies the least squares estimate is presented. By using elimination transformation, an algorithm is developed to solve the models.

关 键 词：线性回归方程组 最小二乘解 算法 线性约束 

分 类 号：O212.1[理学—概率论与数理统计]