I have a question about the calculation of source signals Y and Z

[1DOSAMA]

During the decomposition process, for the first pair, Y(:,1) = X * wt1. As the iteration proceeds to the second pair, Y(:,2) = (X - Driving1) * wt2. So, why is it that after the final decomposition, Y = What.' * X; Z = Vhat.' * X; without taking into account the removal of the driving signal from X during the process?

[dmochow]

Good question -- the rationale for the current design is that the components that are desired reside in the original (non-deflated) data. The removal of the driving component before the second iteration is aimed at identifying a distinct pair of components.