Cannot split data for train, validation and test set in neural network

5 次查看(过去 30 天)
Hello, I want to train a recurrent neural network but I only found one example from the document, and my problem is the training did not split the data set for training, validation and test set even I excuted the code for splitting as the following:
%data conversion
X = con2seq(X);
T = num2cell(T);
%configure RNN
net = layrecnet(1:2,3:5);%3 hidden layer,2 time step
[Xs,Xi,Ai,Ts] = preparets(net,X,T);% prepare input to get time series for training
%training, validation, test split
net.divideParam.trainRatio = 70/100;
net.divideParam.valRatio = 15/100;
net.divideParam.testRatio = 15/100;
net = train(net,Xs,Ts,Xi,Ai);% training the net
view(net);
Y = net(Xs,Xi,Ai);% get prediction
perf = perform(net,Y,Ts);% show the performance
Please tell me where the problem is.
Thanks,
Eason
  1 个评论
Abel Babu
Abel Babu 2017-5-31
Hi Eason,
The above method code is indeed the right way of splitting data. To get a perspective on the question, what make you think the network is not splitting the data?
Abel

请先登录,再进行评论。

回答(0 个)

类别

Help CenterFile Exchange 中查找有关 Deep Learning Toolbox 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