www.gusucode.com > mbcmodels 工具箱 matlab 源码程序 > mbcmodels/@xreginterprbf/evalpev.m
function PEV= evalpev(x,m,FindMax,YVAR); % xreglinear/EVALPEV evaluate Prediction Error Variance for model % % PEV = evalpev(x,m,FindMax) % x is in coded units % m is the model. InitStore must be called on m before this function % FindMax==1 returns -PEV for use in maximising PEV (i.e. G optimality) % % If y data is available % PEV = x'* s^2*inv(X'*X) * x % otherwise PEV = x'* inv(X'*X) * x % % See also xreglinear/PEVGRID % Copyright 2000-2004 The MathWorks, Inc. and Ford Global Technologies, Inc. PEV= zeros(size(x,1),1);