我想在matlab中使用tunelssvm设置LSSVM超参数的初始值,如下所示:
dt = xlsread('USD_GBP.xlsx');
dt1 = dt(:,[1]); %input
x = dt1(1:459,:); %training
x1 = dt1(460:510,:); %testing
dt2=dt(:,[2]); %output
y=dt2(1:459,:); %training
y1=dt2(460:510,:); %testing
[gam, sig2, cost] = tunelssvm({x,y,'f',1,1,'RBF_kernel','preprocess'},'gridsearch', 'crossvalidatelssvm', {10,'mse'})
在代码中,我希望gam和sig2的起始值为1.但是,我得到的sig2的结果是0.005 ...任何人都可以帮助我如何设置起始值或使它们在范围内?