线性约束和奇异SS＆CP矩阵下LS回归的扫描算子

描述在线性约束条件下，且平方和与交叉乘积矩阵在奇异病态情况下，最小二乘回归扫描算子的一些特性。尺管有其他许多不同的方法，由于扫描算法的简单，以及其有关运算矩阵元素所具有的统计意义。

SWEEP OPERATOR FOR LEAST SQUARE REGRESSION SUBJECT TO LINEAR CONSTRAINTS AND SINGULAR SS&CP MATRIX

This paper presents some properties of the SWEEP operator in the scope of least square regression subject to linear constraints and matrix of sum of squares and cross product (SS&CP) in singular ill conditions. Despite alternative approaches available, the SWEEP algorithm represents one of the most flexible tools due to its simplicity in calculation and statiistical implication of the matrix elements involved and, still more importantly, the output interpretation keeps full consistency with the explanation of general output.