MathGroup Archive 2003

[Date Index] [Thread Index] [Author Index]

Search the Archive

activation function for neural network

  • To: mathgroup at smc.vnet.net
  • Subject: [mg42882] activation function for neural network
  • From: "Bruyndonckx P." <pbruynd at vub.ac.be>
  • Date: Fri, 1 Aug 2003 01:25:52 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

I am using the Mathematica neural network package to analyse data.  I want to build a neural network with 3 hidden layers.  I want the neurons in the input layer,
2nd hidden layer and the output layer to be linear.  The neurons in the first and third hidden layer should have a sigmoidal activation function.  I haven't found
a way to change the activation function in certain layers only.  It is possible to change the activation from linear to sigmoidal in the output layer, and it is
also possible to change to activation function to any function for ALL layers.  Can somebody help me on this problem.

Thanks,

Peter


  • Prev by Date: Severe ToExpression Fault?
  • Next by Date: multiple roots finder.
  • Previous by thread: Re: Severe ToExpression Fault?
  • Next by thread: multiple roots finder.