matthiasm@8: function [PCcoeff, PCvec] = pca(data, N)
matthiasm@8: %PCA	Principal Components Analysis
matthiasm@8: %
matthiasm@8: %	Description
matthiasm@8: %	 PCCOEFF = PCA(DATA) computes the eigenvalues of the covariance
matthiasm@8: %	matrix of the dataset DATA and returns them as PCCOEFF.  These
matthiasm@8: %	coefficients give the variance of DATA along the corresponding
matthiasm@8: %	principal components.
matthiasm@8: %
matthiasm@8: %	PCCOEFF = PCA(DATA, N) returns the largest N eigenvalues.
matthiasm@8: %
matthiasm@8: %	[PCCOEFF, PCVEC] = PCA(DATA) returns the principal components as well
matthiasm@8: %	as the coefficients.  This is considerably more computationally
matthiasm@8: %	demanding than just computing the eigenvalues.
matthiasm@8: %
matthiasm@8: %	See also
matthiasm@8: %	EIGDEC, GTMINIT, PPCA
matthiasm@8: %
matthiasm@8: 
matthiasm@8: %	Copyright (c) Ian T Nabney (1996-2001)
matthiasm@8: 
matthiasm@8: if nargin == 1
matthiasm@8:    N = size(data, 2);
matthiasm@8: end
matthiasm@8: 
matthiasm@8: if nargout == 1
matthiasm@8:    evals_only = logical(1);
matthiasm@8: else
matthiasm@8:    evals_only = logical(0);
matthiasm@8: end
matthiasm@8: 
matthiasm@8: if N ~= round(N) | N < 1 | N > size(data, 2)
matthiasm@8:    error('Number of PCs must be integer, >0, < dim');
matthiasm@8: end
matthiasm@8: 
matthiasm@8: % Find the sorted eigenvalues of the data covariance matrix
matthiasm@8: if evals_only
matthiasm@8:    PCcoeff = eigdec(cov(data), N);
matthiasm@8: else
matthiasm@8:   [PCcoeff, PCvec] = eigdec(cov(data), N);
matthiasm@8: end
matthiasm@8: