ABSTRACT
Our goal in this chapter is modest, of necessity. Having spent a book’s worth in the exposition of MRC, obviously several books would be necessary for com parable exposition of multivariate (MV) methods. All we mean to do here is sketch out the relationship of MRC to these, provide some information about their purposes and characteristics, and show how MRC can accomplish some of these and, in some circumstances, more. In doing this, we show that the MRC method, as general as we have demonstrated it to be, is itself a special case of the yet more general method of canonical correlation and regression analysis (CA), which (inverting the relationship) not only specializes to MRC but also to the other major MV methods. The mathematical structure is inevitably quite compli cated-about an order of magnitude more so than MRC.
