ABSTRACT
As pointed out in Chapter 2 , matrix algebra proves extremely useful in multivariate statistics for at least three reasons:
Matrix notation provides a highly compact summary of expressions and equations that, if a separate symbol for each variable and each constant involved were employed, would be enormously bulky.
The compactness and simplicity of matrix notation greatly facilitates memorization of expressions and equations in matrix form.
The underlying similarity of the operations involved in applying a multivariate technique to a set of 2, 15, or 192 variables is much easier to see when these operations are expressed in matrix notation. Further, the similarity between univariate and multivariate techniques is much more easily seen from matrix expressions than from single-symbol formulas. These two functions combine to lessen considerably the difficulty of deriving solutions to multivariate problems.
