0
$\begingroup$

I'm modeling markov chains for some data sets.But not all of them has unique stationary state.

I create the transaction matrix A from the data sets. If Ahas no unique stationary state, How can I change A to get a proper approximate unique stationary state for my markov chain modeled from the data set?

  • 0
    I've read your question, but am unsure exactly what you're asking. Consider revising your question with more detail and accuracy. For instance, by "transaction matrix" do you mean "transition matrix?" What exactly do you mean by "stationary state?" Positive recurrent perhaps? Perhaps you mean a single entry in $A$ that is identical in $A^k$ for any $k$? – 2017-02-07
  • 0
    Is $A$ a stochastic matrix? – 2017-02-09

0 Answers 0