How to calculate network value with 2 input vectors .

This is a very basic code to know how NARX Network works . d1=[1:2];

    d2=[1:2];
    net = narxnet(d1,d2);
    net.inputs{1}.size=2;
    net.inputs{2}.size=1;
    net.numLayers=1;
    net.biasConnect=0;
    net.layers{1}.transferFcn='purelin';
    net.layers{1}.dimensions=2;
    net.IW{1,1}=[1 2 2 1; 3 1 1 4];
    net.IW{1,2}=[1 3; 4 5];
    x={[1 2] [3 4] [5 6] [7 8] };
    y={10 12 16 18};
    A=net(x,y)
This gives error 
Error using network/sim (line 280)
Number of inputs does not match net.numInputs.

Error in network/subsref (line 17) otherwise, v = sim(vin,subs{:}); I have tried to understand Narx network in the same way as in the "neural network toolbox's user guide" explained about "Simulation with concurrent inputs in a dynamic network". Where is the error

[Greg Heath]

Is there any reason why you won't use the help and doc example as a guide?

Answer 1

[Greg Heath]

Is there any reason why you won't use the help and doc example as a guide?

For example, run the example. Then look at the net's properties to understand what was created.

Your x and y are not compatible

>> X = cell2mat(x), Y=cell2mat(y), whos X x Y y

X =

     1     2     3     4     5     6     7     8

Y =

    10    12    16    18

Name      Size            Bytes  Class     Attributes

X         1x8                64  double              
Y         1x4                32  double              
x         1x4               304  cell                
y         1x4               272  cell

Did you mean

x={[1; 2] [3; 4] [5; 6] [7; 8] };

Hope this helps.

Thank you for formally accepting my answer

Greg

[sandeep]

Sir, Example on NARX network is on real problem, This type of small example is not available so that I can check manually how it works.

x={[1; 2] [3; 4] [5; 6] [7; 8] }; I also tried with this, but it was not working.

[Greg Heath]

What are you thinking? How is the response "But it was not working" supposed to help me help you? Did you get an error message or just the answer you don't want?

I repeat: Try a MATLAB data set. Once it runs, you can truncate the data and investigate a small problem.

help nndata