Second Order Sobol Indices

4 次查看(过去 30 天)
Ian Loveless
Ian Loveless 2018-7-5
I am trying to produce second order Sobol indices for feature selection. I am looking for the features that best predict the binary outcome variable. I keep getting an error saying that my iteration reached because the estimated coefficients perfectly separate failures from successes.
Here is my code:
for i=1:1000
Mdl=fitglm([xt(:,i) xt(:,i).^2],yt,'distribution','binomial');
dyhat=predict(Mdl,[xt(:,i) xt(:,i).^2]);
dyhat=dyhat>0.5;
te(:,i)=sum(dyhat(1:n0)>0)+sum(dyhat(101:end)<1);
[~,I]=sort(te);
end
Are there obvious bugs that I am missing? Thank you for any help.

回答(0 个)

类别

Help CenterFile Exchange 中查找有关 3-D Scene Control 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by