2

One of the main reasons that ANN performs better is due to its influential feature in handling nonlinear data (Wu et al, 2008).

Can anyone explain to me what is this meaning?

bbadyalina
  • 743
  • 2
  • 7
  • 20

1 Answers1

1

This means that a neural network can have activation functions that aren't linear, which causes the neural network's output to be non-linear with respect to the input.

For further information, this blog post gives some nice illustrations and explanations: http://colah.github.io/posts/2014-03-NN-Manifolds-Topology/

Franck Dernoncourt
  • 42,093
  • 30
  • 155
  • 271
  • The question is about why ANN can handle non-linear data so well, and this is not really explained in your answer. – Pieter Jan 20 '17 at 07:54
  • 1
    @Pieter I interpreted the question how comes neural networks can handle nonlinear data. I agree it could be interesting to explain why ANN can handle non-linear data so well though, but that is more challenging :-) – Franck Dernoncourt Jan 20 '17 at 16:04