0

Are there any common ways to define a distance between two probablity distribution functions (i.e. how similar two distributions are)?

I am interested in the specific case where the random variable is discrete and can only take a finite number of values.

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Tony
  • 101
  • A simple [search](http://stats.stackexchange.com/search?q=distance+distributions) on this site gives you lots of related questions. – Richard Hardy Oct 15 '15 at 10:49

1 Answers1

1

The Kullback-Leibler divergence (i.e. relative entropy) does exactly that, as long as the two distributions have the same support.

ocramz
  • 331
  • 1
  • 6