Plotperformance of several neural networks
1 view (last 30 days)
Show older comments
Dmitry Devjatykh
on 23 Nov 2015
Commented: Greg Heath
on 23 Nov 2015
Hi. I want to compare performance of neural network depending on training algorithm.
[net,tr] = train(net,inputs,targets,inputStates,layerStates);
[net1,tr1] = train(net1,inputs,targets,inputStates,layerStates);
Now the problem is that when I do
plotperform(tr,tr1) %%first variant
plotperform(tr,t,tr1,t) %%sendond variant
I get errors. If I do
plot(tr.perf,t,tr1.perf,t)
I get plots that are very undetalized, they almost look like straight lines. So my question is is it possible to use plotperf function to plot several neural nets at a time? or at least how to increase detalization of plot function, because it simplifies performance curve too much compared to plotperf.
0 Comments
Accepted Answer
Greg Heath
on 23 Nov 2015
plotperform only takes one tr as input.
You can use plot. However, you have to provide the details yourself.
To plot multiple curves you have to provide input and output for each curve
plot(x1,y1,x2,y2,... )
Hope this helps.
Greg
2 Comments
Greg Heath
on 23 Nov 2015
Yes, it is. However, I cannot help with that. Your best bet is to look at the source code via
type plotperform
Greg
More Answers (0)
See Also
Categories
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!