Neural network input -output problem

I am new to the Neural network toolbox. I want to do a simple experiment with the neural network fitting tool.

I have a file called datain which contains a single column matrix of 9 elements. I have a target which is a file called dataout that is a single column matrix of 9 elements with similar values to datain. I can load datain and dataout using the fitting tool GUI. However after training I get the message

"the input data has more than one element The fitting tool can only plot single input problems"

My guess is that I have got muddled somewhere here. Can anyone explain what is going wrong?

Answer 1

[Christopher Bishop]

Thanks Greg. It all appears to work now. It seems that matrix inputs to the NNFT in MATLAB have to be transposed from columns as you say. My primary data was in an Excel spreadsheet as a simple column. I transposed the data using the transpose option on Excel 'pastespecial' and saved it as a new file. The NNFT accepts this now without problem.

[Greg Heath]

You are welcome. However, I would appreciate you accepting my answer instead of your response to my answer which should be moved to a comment box.

Answer 2

[Greg Heath]

Try transposing your matrices

Hope this helps

Greg

[Farhana Razak]

Hi there. I'm having the same problem here and I already transposed my data but the problem still occured. Do you have any idea why is this happens?