I am empirically creating Markov chains similar to this question. I end up with several finite state Markov chain models with the same nodes but varying transition probabilities. I want to calculate a pairwise distance between those models. What are appropriate distance measures to compare two Markov chains?
In my case the state space are user actions in a computer system and the transition probabilities are the observed transition frequencies between different actions.
Best Answer
As the problem comes down to compare two probability vectors, I guess you could use for instance the Kullback–Leibler divergence or, if you need a symmetrical measure, the Jensen–Shannon divergence.
Hope this help