How can I reuse the same neural network to recreate the same results I had while training/creating the network?

Hello,

I am new to neural networking, so my question may sound stupid.

 I have trained/created a neural network, I have saved the script of the network, but each time I run the script, it gives different R values. How can I get the same R value I had while training the network and how do I plug in new input data to see the results by using the same network.

Answer 1

[Greg Heath]

I don't think you want to save the script.

You want to save the net

save net

Hope this helps.

Greg

Answer 2

[the cyclist]

Presumably you need to set the random number generator seed.

If you have a relatively new version of MATLAB, you can do this with the rng() command, for example, put

rng(1)

at the beginning of your code.

doc rng

for details.

[Karthik]

Thanks for the fast reply, it works:)

But, when I run the network, with through ''nnstart'' and train it, it gives an R value of 0.98 , but once I save the script and rut it its gives 0.60 as R value, each time I run, not 0.98. Also how Can i use the saved network for new testing new datas.

[Greg Heath]

1. (0.6 vs 0.98): Need more info

2. ynew =net(xnew);

Answer 3

[Karthik]

I fixed the first problem. _ | *Regarding the 2nd question, I trained a network, then I saved the script as 'ntwk'. Then I went to my matlab file, which has the data that was used for training, and created a new array called 'xnew', to check how my saved network works on the new data, then i tried to execute the command _ .|*

ynew=ntwk(xnew)

IT GIVES THE ERROR MESSAGE

'' Attempt to execute SCRIPT ntwk as a function: H:\NEURAL NETWORK\ntwk.m

Error in valmetrutpercent (line 112) ynew=ntwk(xnew) ''

THE NETWORK CODE I SAVED WAS

% Solve an Input-Output Fitting problem with a Neural Network % Script generated by NFTOOL % Created Thu Mar 06 18:11:55 CET 2014 % % This script assumes these variables are defined: % % ci10_1 - input data. % z10_1 - target data.

rng(1)

inputs = ci10_1; targets = z10_1;

% Create a Fitting Network hiddenLayerSize = 10; net = fitnet(hiddenLayerSize);

% Setup Division of Data for Training, Validation, Testing net.divideParam.trainRatio = 70/100; net.divideParam.valRatio = 15/100; net.divideParam.testRatio = 15/100;

% Train the Network [net,tr] = train(net,inputs,targets);

% Test the Network outputs = net(inputs); errors = gsubtract(targets,outputs); performance = perform(net,targets,outputs)

% View the Network view(net)

% Plots % Uncomment these lines to enable various plots. %figure, plotperform(tr) %figure, plottrainstate(tr) %figure, plotfit(net,inputs,targets) %figure, plotregression(targets,outputs) %figure, ploterrhist(errors)

[Greg Heath]

Please do not use the ANSWER box for comments. In this case

1. copy your comments in the answer bos

2. paste them in the comment box

3. delete your answer box