00001 function [ssx, y2] = dofit(y, Qd) 00002 %DOFIT Do constrained least squares fit and reduce data for subsequent fits 00003 00004 % Fit y to design matrix in null space 00005 y2 = Qd' * y; % rotate y into that space: predicted y value 00006 ssx = norm(y2)^2; % sum of squares explained by fit
This site conforms to the following standards: