total sum of square

In this paper,the sum of squares of an a-level factor is decomposed to a-1 sums of squares of mutually orthogonal contrasts,such that the total variation sum of squares is decomposed to a parts including the residual sum of squares.

本文把a水平因子的平方和分解成相互正交的a-1个对照的平方和,这样总变差平方和就可以分解成a个部分(包括残差项),然后又将该分解方法推广到了多因子的情形,并通过因子平方和的分解找到了多因子交互效应对应的对照向量,这使得多水平因子交互效应的计算和解释更加容易,也为方差分析带来了更多的方便,最后给出了几个应用示例。