MATLAB实现的正交信号校正OSC源代码解析

OSC源码
osccalc.m for PC  
function [x,nw,np,nt] = osccalc(x,y,nocomp,iter,tol)
%OSCCALC Calculates orthogonal signal correction
%  The inputs are the matrix of predictor variables (x)
%  and predicted variable(s) (y), scaled as desired, and
%  the number of OSC components to calculate (nocomp).
%  Optional input variables are the maximum number of
%  iterations used in attempting to maximize the variance
%  captured by orthogonal component (iter, default = 0),
%  and the tolerance on percent of x variance to consider
%  in formation of the final w vector (tol, default = 99.9).
%  The outputs are the OSC corrected x matrix (nx) and
%  the weights (nw), loads (np) and scores(nt) that were
%  used in making the correction. Once the calibration is 
%  done, new (scaled) x data can be corrected by 
%  newx = x - x*nw*inv(np'*nw)*np';
%
%I/O: [nx,nw,np,nt] = osccalc(x,y,nocomp,iter,tol);
%
%See also: CROSSVAL

%Copyright Eigenvector Research, Inc. 1998
%Barry M. Wise, January 23, 1998

[m,n] = size(x);
nw = zeros(n,nocomp);
np = zeros(n,nocomp);
nt = zeros(m,nocomp);
if nargin < 4 | isempty(iter)