cross validation for neural network

2 次查看(过去 30 天)
i want to use cross validation method to decide the number of hidden neurons of a neural network.
i want 5 fold cross validation. and right now i am using following NN architecture:
if true
net=newff(minmax(in'),[7,3],{'tansig','purelin'},'traingdx');
net.trainParam.show = 50;
net.trainParam.lr = 0.05;
net.trainParam.mc = 0.7;
net.trainParam.epochs = 3500;
net.trainParam.goal = 1e-2;
a1 = net.b{1};
a2 = net.b{2};
w1 = net.iw{1};
w2 = net.lw{2};
end
how can i use cross validation for this. and where the errors of each fold will be stored......

采纳的回答

Greg Heath
Greg Heath 2014-2-12
Search the NEWSGROUP and ANSWERS using
greg crossvalidation
and
greg cross-validation
and
greg cross validation
Please post the addresses of any posts that are useful.
Hope this helps.
Thank you for formally accepting my answer
Greg

更多回答(0 个)

Community Treasure Hunt

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

Start Hunting!

Translated by