Validation and train sets are equal?
信息
此问题已关闭。 请重新打开它进行编辑或回答。
显示 更早的评论
When I train this net:
net = feedforwardnet(20,'trainlm');
net.inputs{1}.processFcns = {};
net.outputs{2}.processFcns = {};
net.trainParam.epochs = 200;
It looks like, that I have the exactly same train and validation sets during training, so it never ends until max epoch is reached. I have tried to change ratio manually, but it doesnt work as well.
My input data set is unique.
I've tried another input data set and it works. I have also tried to specific range for validation and train set by divideint and it works as well:
net.divideFcn = 'divideind';
net.divideParam.trainInd = 89:584;
net.divideParam.valInd = 1:88;
So what am I missing?
1 个评论
Greg Heath
2016-8-27
It would help if you posted your commented code operating on one of the MATLAB example data sets.
help nndatasets
doc nndatasets
Greg
PS: What is wrong with using the example code in
help feedforwardnet
doc feedforwardnet
?
回答（0 个）
此问题已关闭。
关闭：
2021-8-20
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
