MATLAB: Problem with configuration of neural network

neural networktrain

hi I have a problem with using neural network. although I set all the parameter in 'trainparam' but it uses its default setting in training. the method that i set the parameter is as below: if true %
end
L=[4 8];
net=newcf(P,T,L);
net.trainparam.goal=1e-5;
net.trainParam.min_grad=1e-5;
net.trainparam.epochs=200;
net.trainFcn='traincgf';
net.layers{1}.transferFcn='tansig';
net.layers{2}.transferFcn='tansig';
[net,tr]=train(net,P,T);

Best Answer

>> help newcf
   Create a cascade-forward backpropagation network.
   Obsoleted in R2010b NNET 7.0.  Last used in R2010a NNET 6.0.4.
   The recommended function is cascadeforwardnet.
Using the data in help and rng(0) to initialize the random number generator, I get
    close all, clear all, clc
    P = [0 1 2 3 4 5 6 7 8 9 10];
    T = [0 1 2 3 4 3 2 1 2 3 4];
    L=[4 8];
    net=newcf(P,T,L);
%Defaults
    net.trainparam.goal         % 0
    net.trainParam.min_grad     % 1e-5
    net.trainparam.epochs       % 1000
    net.trainFcn                % trainlm
    net.layers{1}.transferFcn   % tansig



    net.layers{2}.transferFcn   % tansig
%Assigned
    net.trainparam.goal        = 1e-5;
    net.trainParam.min_grad    = 1e-5;
    net.trainparam.epochs      = 200;
    net.trainFcn               = 'traincgf';
    net.layers{1}.transferFcn  = 'tansig';
    net.layers{2}.transferFcn  = 'tansig';
    rng(0)
    [net,tr,Y]=train(net,P,T);
    NMSE = mse(T-Y)/var(T,1)    % 0.2933
%Final
    net.trainparam.goal         % => 0
    net.trainParam.min_grad     % => 1e-10
    net.trainparam.epochs       % => 1000
    net.trainFcn                % traincgf
    net.layers{1}.transferFcn   % tansig
    net.layers{2}.transferFcn   % tansig
Which shows that goal, min_grad, and epochs have changed
However, only goal and epochs have changed to default values
Obviously, there are bugs.
The important thing is whether the normalized MSE is acceptable when longer data sets are used
Hope this helps.
Thank you for formally accepting my answer
Greg
Related Question