Kohonen (example from help does not work)

1 view (last 30 days)
Hi all, please I need your help !
example (Kohonen) from help does not work
>> p = [.1 .8 .1 .9; .2 .9 .1 .8];
>> net = newc([0 1; 0 1],2);
>> wts = net.IW{1,1}
>> biases = net.b{1};
>> net.trainParam.epochs = 500;
>> net = train(net,p);
>> a = sim(net,p);
>> ac = vec2ind(a); >> ac
ac =
1 1 1 1
>> wts
wts =
0.5000 0.5000
0.5000 0.5000
My weight did not change. in example
wts =
0.1000 0.1467
0.8474 0.8525
Maybe I'm doing something wrong?

Accepted Answer

Alexander
Alexander on 8 Sep 2011
7.10.0 (R.2010b)
  1 Comment
Lucas García
Lucas García on 8 Sep 2011
So, download the fix in the previous link and give it a try. Then it will work fine.

Sign in to comment.

More Answers (3)

Walter Roberson
Walter Roberson on 7 Sep 2011
You did not assign a new value to wts, so it is going to continue to be the value that was retrieved before you did the training. You need to retrieve net.IW{1,1} after the training in order to see the effect of the training on the weights.
  1 Comment
Alexander
Alexander on 8 Sep 2011
These are my values after training
net.IW{1,1}
ans =
0.5000 0.5000
0.5000 0.5000
After training, they have not changed.

Sign in to comment.


Lucas García
Lucas García on 8 Sep 2011
Which release are you using? There is a known bug in R2010b and R2011a regarding newc and competlayer.
It has been fixed in R2011b.

Alexander
Alexander on 8 Sep 2011
It is strange, but FIX_R2010b_COMPETLAYERAYER.zip is corrupted.
FIX_R2011a_COMPETLAYER.zip is normal.

Categories

Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!