How can I plot a neural network?

33 Ansichten (letzte 30 Tage)
Franck Dernoncourt
Franck Dernoncourt am 22 Jan. 2013
Beantwortet: Sajjad Shawl am 5 Jun. 2022
Say I created a neural network, such as net = feedforwardnet(30);, which I trained [net tr] = train(net,p,t);
How can I plot this neural network? I would like something like:
(I'm especially interested in having the weights displayed on the graph)
Thanks, Franck

Akzeptierte Antwort

Greg Heath
Greg Heath am 23 Jan. 2013
The only NNTBX display that is available is
view(net)
However, perhaps you can find something online.
Hope this helps.
Thank you for accepting my answer.
Greg
  1 Kommentar
Franck Dernoncourt
Franck Dernoncourt am 23 Jan. 2013
Thanks for your answer, view() was indeed the only function I could find… But I'm surprised there is no existing code for such a common task. Maybe some function to output the network to a .dot file that we would give to GraphViz?

Melden Sie sich an, um zu kommentieren.

Weitere Antworten (3)

Yashwant Kolluru
Yashwant Kolluru am 12 Jul. 2018
[a,b] = train(net,x,t); y = net(x); e = gsubtract(t,y); %%%%%%%%%%%%%%%%%%% figure, plotperform(b); figure, plottrainstate(b); figure, ploterrhist(e); figure, plotregression(t,y);
hope it helps

fred  ssemwogerere
fred ssemwogerere am 2 Jan. 2020
Hello, with the latest Matlab version 2019b, you are availed with a wide range of functions that can do this job for you. With my program for neural network modeling, i added some lines of code that also plot the best neural network. You can start by checking out the "line" and "viscircles" functions in the Matlab documentation. These can help you with what you want to plot. Below is a sample plot output from my code, of the best network i was designing. With added work, you can modify the lines to visualize the magnitude and direction of the weights, remove the axis bars, and many other things that you would like to do.
Regards,
Fred
neural.png
  1 Kommentar
Maha Mosalam
Maha Mosalam am 1 Mär. 2021
can you submit the line codes to do this, please

Melden Sie sich an, um zu kommentieren.


Sajjad Shawl
Sajjad Shawl am 5 Jun. 2022
hi
have a look at this code from matlab exchange
quite usefull
Call the function to plot FNN: plotFNN(layersLength), where layersLength is a 1*n row vector containing the number of neurons in each layer.
https://www.mathworks.com/matlabcentral/fileexchange/87287-plot-feedforward-neural-networks), MATLAB Central File Exchange. Retrieved June 5, 2022.

Kategorien

Mehr zu Sequence and Numeric Feature Data Workflows finden Sie in Help Center und File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by