different result is obtained for the same input using self-organizing map

1 visualizzazione (ultimi 30 giorni)
liangjian
liangjian il 28 Feb 2012
Commentato: Mahmoud ELKAFAFY il 14 Ott 2019
I used Neural Network Toolbox to train a self-organizing map for a given data set. I found that each run gave me a quite different result, like different "sample hits map". How to understand this difference? I just followed the standard routine, like
net = selforgmap([8 8]);
net = train(net, testInput);

Accedi per rispondere a questa domanda.

Risposte (1)

Greg Heath
Greg Heath il 3 Mar 2012
Initialize the random number generator to the same intial state; otherwise the results will be different.
Hope this helps.
Greg
  1 Commento
Mahmoud ELKAFAFY
Mahmoud ELKAFAFY il 14 Ott 2019
But, if you change the initialization of random number generator (e.g., rng(7) instaed of rng(3) for insatance), you will still get different results.

Accedi per commentare.

Categorie

Scopri di più su Function Approximation, Clustering, and Control in Help Center e File Exchange

Tag

Prodotti

Community Treasure Hunt

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

Start Hunting!

Translated by