Statistical Inference for Markov chains based on divergence measures
In the present work we propose a class of test statistics based on the family of weighted φ-divergences for general order Markov chains. A weight matrix treats the issue of the presence (or not) of prior information on the transitions of the system. That methodology could be adapted in the framework of homogeneity or goodness-of-fit for Markov chains. The appropriate asymptotic theory is presented according with Monte Carlo simulations for assessing the performance of the proposed test statistics.