0

So I'm working on some simple neural network problems for a class I'm taking, and I've mostly done all of the work.

However, I'd like to be able to press run and have all of the relevant figures and graphs pop out so I can easily show my work.

To do this I've been drawing setting the names of figures that I plot in using the

figure('Name', 'NameGoesHere', 'NumberTitle', 'off')`

variant of the figure function.

The thing is, one of the things that I have to show are performance graphs for the different networks I made, and using plotperform(tr) sets the window name to Performace (plotperform) even if I try setting the figure name to something else. Any help?

Here's a picture showing what I mean:

Screenshot

Oh, and, being really new to Matlab, correct me if I'm wrong, but since performance is defined as mean absolute error, the lower it is, the better right? So is there a way to set is so plotperform actually selects the lowest point of the graph as the best instead of the first one?

Another thing I'd like to be able to do is have the training information nntraintool open in a new window every time I call the train(netA, P, T) function. Right now, every time it's called it just overwrites the previous info in the nntraintool window. Is it even possible to do this?

This is the window I'm talking about:

nntraintool window

mkierc
  • 1,193
  • 2
  • 15
  • 28
VelikiDrMr
  • 1
  • 1
  • Can you please post the code for what you've tried, and the error code or undesired behavior it generates? – darthbith Jan 18 '14 at 03:51
  • There's no error or undesired behaviour. As I said, I just want to make a few aesthetic changes and was wondering if Matlab allows such a thing. – VelikiDrMr Jan 18 '14 at 06:39
  • "using plotperform(tr) sets the window name to...even if I try setting the figure name to..." This seems like undesired behavior. Can you please post your code? It seems like your first question is answered here: http://stackoverflow.com/a/4684383/2449192 Your second question is very difficult to answer without seeing the code, and the third doesn't seem possible based on a cursory Google search, but I have very little experience with the Neural toolbox. Hope it helps! :-) – darthbith Jan 18 '14 at 14:18

0 Answers0