Neural Network problem with the goal

1 visualización (últimos 30 días)
Andres  Romero
Andres Romero el 1 de Oct. de 2016
Respondida: Greg Heath el 2 de Oct. de 2016
Hey everybody maybe someone can help me, with this. What I could do if I put this value in the goal of the net, and the net can't get to this value and stop in validation check net.trainParam.goal=0.01*mean(var(target',1)) What can I do? or I am wrong doing this... I'm a bit lost. Thanks for help me...

Respuestas (1)

Greg Heath
Greg Heath el 2 de Oct. de 2016
For difficult problems, success beyond the default parameter values typically depends on the number of hidden nodes and choice of initial random weights.
Typically I design 100 nets at a time in a double loop configuration with an outer loop of 10 values of hidden nodes Hmin:dH:Hmax and an inner loop with Ntrials = 10 of random initial weights.
Search both NEWSGROUP and ANWSERS for some of my examples using
greg Hmin:dH:Hmax Ntrials
Hope tis helps
Thank you for formally accepting my answer
Greg

Categorías

Más información sobre Sequence and Numeric Feature Data Workflows en Help Center y File Exchange.

Etiquetas

Community Treasure Hunt

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

Start Hunting!

Translated by