0

I am using PRtools back-propagation based feed forward neural network classifier (e.g.[W, HIST]=bpxnc(Dataset,5,inf,[])). I am trying to figure out if we can specify how many nodes each layer has or if there is a way to specify the number of nodes for my classifier?

Thanks for help, Sabin

Sabin
  • 1

1 Answers1

0

Check PRTools manual. 

[W,HIST] = BPXNC (A,UNITS,ITER,W_INI,T,FID)

UNITS Array indicating number of units in each hidden layer (default: [5])

Chris
  • 460
  • 2
  • 12
  • Thanks Chris. But I was wondering about number of nodes in each layers. – Sabin Nov 25 '13 at 18:30
  • Based on the description of BPXNC in the PRTools manual I guess UNITS is a vector. The length of the vector determines the number of hidden layers and each value of the vector determines the number of nodes on the corresponding hidden layer. – Chris Nov 26 '13 at 22:51