nobias.m

SVR程序,直接用就可以了,没有错误,其中还有一些变量需再编程序算,如:MSE

function nb = nobias(ker)
%NOBIAS returns true if SVM kernel has no implicit bias
%
%  Usage: nb = nobias(ker)
%
%  Parameters: ker - kernel type
%
%  Author: Steve Gunn ([e-mail])

if (nargin ~= 1) % check correct number of arguments
    help nobias
else

    switch lower(ker)
        case {'linear','sigmoid'}
            %,'anovaspline1','anovaspline2','anovaspline3'}
            nb = 1;
        otherwise
            nb = 0;
    end

end