www.gusucode.com > mbcmodels 工具箱 matlab 源码程序 > mbcmodels/@xregmodel/cov.m
function c = cov(m) %COV Covariance matrix % % C = COV(M) % % The covariance is given by cov = Ri*Ri' where ri is the information matrix % and is given by Ri = var( m ) = or by Ri*Ri' = inv( X'*X ) * mse. Thus % cov( m ) = inv( X'*X ) * mse, where mse is the mean square error and X is the % Jacobian of the model. % % As we don't know the MSE we assume that it is 1 as this should only be used % by designs. % % The information matrix needs to be stored in the model via a call to % VAR(M,...). This is usually done by the INITSTORE method. % % See also XREGLINEAR/COV, XREGUSERMOD/CALCJACOB, XREGMODEL/INITSTORE, % XREGMODEL/VAR. % Copyright 2000-2006 The MathWorks, Inc. and Ford Global Technologies, Inc. % The information matrix is stored in the model. This can be accessed by a call % to var. ri = var( m ); if ~isempty(ri) % Compute the variance. c = ri*ri'; else % default covariance c= zeros(numParams(m)); end %------------------------------------------------------------------------------| % EOF %-------------------------------------------------------------------------- %----|