MATLAB: How to customize the NARX neural network Input and Feedback Pre/Post-Processing Functions

mapminmaxnarxneural networksprocessfcns

Hi,

I am new to neural networks. I have 1 target and 4 external inputs and I need to customize the NARX neural network Input and Feedback Pre/Post-Processing Functions so as to map onto the range [0.2 0.8] instead of the default [-1 1]. How can I modify these two lines in my script, and what else can I do?

net.inputs{1}.processFcns = {'removeconstantrows','mapminmax'};

net.inputs{2}.processFcns = {'removeconstantrows','mapminmax'};

Thanks.

Best Answer

Why do you thin you "need" to do this transformation??

MAPMINMAX is designed to scale internal inputs, targets and outputs into a range that doesn't unnecessarily cause numerical innaccuracy in training function Jacobians or saturate sigmoid activation functions.

Changing from [-1,1] to something else does absolutely nothing useful.

The change of code is an unecessary complication that does nothing to improve the accuracy of the net.

Or am I missing something?

Hope this helps.

Thank you for formally accepting my answer (;>)

Greg

Related Solutions

MATLAB: Change preprocessing parameter in neural network

You can try this workaround:

net = struct(net);
net.inputs{1}.processParams{2}.ymin = 0.1;
net = network(net);

MATLAB: Training a neural network

Sam harris on 2 Jul 2012

 % Create a Nonlinear Autoregressive Network with External Input
 % inputDelays = 1:1; feedbackDelays = 1:1; hiddenLayerSize = 10;

1. What makes you think these are appropriate inputs??

 % net =narxnet(inputDelays,feedbackDelays,hiddenLayerSize); 
 % net.inputs{1}.processFcns = {'removeconstantrows','mapminmax'};
 % net.inputs{2}.processFcns = {'removeconstantrows','mapminmax'};

2. Why bother? The last 2 statements are defaults.

 % [inputs,inputStates,layerStates,targets] = preparets(net,inputSeries,{},targetSeries); 
 % net.divideFcn = 'dividerand'; % Divide data randomly
 % net.divideMode = 'value'; % Divide up every value

3. The last 2 statements are inappropriate for time series

 % net.divideParam.trainRatio = 70/100; net.divideParam.valRatio = 15/100;
 % net.divideParam.testRatio = 15/100; 
 % net.trainFcn = 'trainlm'; % Levenberg-Marquardt 
 % net.performFcn = 'mse'; % Mean squared error 
 % net.plotFcns = {'plotperform','plottrainstate','plotresponse', ...
 % 'ploterrcorr', 'plotinerrcorr'};

4. Why bother? The last 6 statements are defaults.

 % % Train the Network 
 % [net,tr] =train(net,inputs,targets,inputStates,layerStates);
 % if true % code 
 % end

5. What does "if true ...etc... end" suppose to do?

6. You still have to close the loop and continue training.

7. See

http://www.mathworks.com/matlabcentral/newsreader/view_thread/332147#912806

Best Answer

Related Solutions

MATLAB: Change preprocessing parameter in neural network

MATLAB: Training a neural network

Related Question