neural network error Undefined function or variable 'network'?

Hi all, There is an error while I used neural network toolbox. I wrote a script, but when I start running it I getting

Undefined function or variable 'network'.

And this what I do

 clear
load('Metdata','METData')
training_set=METData(1:5088,:);%specify training set
shuffling_inputs=training_set(randperm(5088),1:4);%randomize the input matrix
load('SLR');
target_set=SLR(1:5088,3);%sea level residual targrt set (1-7)
testing_set=METData(5089:7296,:);%specify testing set (1-7)
testing_target_set=SLR(5089:7296,3);%specify testing set (8-11)
[pn,ps]=mapminmax(training_set');%Normalize the input and output -1,1
[tn,ts]=mapminmax(target_set');

 net = feedforwardnet(10);

how can I solve this problem?

thank you all.

Answer 1

[Greg Heath]

It is clear that you do not understand how the MATLAB toolbox functions work.

 0. MATLAB DOESN'T USE UPPER CASE. I AM USING IT FOR EMPHASIS.
 1. FITNET is used for regression/curve-fitting
 2. PATTERNNET is used for classification/pattern-recognition
 3. Both call FEEDFORWARDNET automatically
 4. INPUT(x) and TARGET(t) sets are used in the function TRAIN.
 5. TRAIN automatically divides the data into training/validation/testing subsets in the 
 ratio 0.7/0.15/0.15
 6. It is advisable to initialize the RNG before training so that your results obtained 
 from automatic random data division and random initial weights can be reproduced.
 7. The syntax
 [net tr y e ] = train(net,x,t);
 yields the OUTPUT y = net(x), ERROR e = t-y,TRAINING RECORD tr and TRAINED NET net. 
 8. I have posted many examples. Search the NEWSGROUP and ANSWERS using

 greg fitnet tutorial

and

 greg fitnet

Hope this helps.

Thank you for formally accepting my answer

Greg

[Lilya]

thank you Dr. greg I want to make sure if the script is correct or not to keep working on it

Thank you again

[Greg Heath]

No it is not correct.

Take a look at my fitnet tutorials.

Post back if you have questions about the tutorials.

Greg

[Lilya]

Dr. Greg I'm looking for nonlinear autoregressive in NN toolbox Did you post a tutorial about it? The information in the doc narnet isn't clear enough for me thank you for your help

[Greg Heath]

Your original post used FEEDFORWARDNET appropriate for regression & classification. Is this the same problem? If not, you should have started another thread with NARNET in the title.

Of course I have tutorial posts on NARNET.

All you had to do was search BOTH the NEWSGROUP and ANSWERS using

 greg narnet

and

 greg narnet tutorial

Although my explicit tutorials (TUTORIAL in the title) are in the NEWSGROUP, search both forums with and without the term tutorial.

Hope this helps.

Greg

[Lilya]

ok now its clear I will study your tutorial perfectly. the data which I work with is atmospheric parameters to estimate the sea level variations (actually the rule of each parameter in the sea level) to predict the time series autoregrissin (NARX) is more convenient to use

Answer 2

[dulaj sanjaya]

I got the same error. I solved it.The reason is you haven't verified license. if you are using linux try to install as root user.

sudo ./install

After successful installation you will be asked to verify license.