Consistency of the maximum likelihood estimator in seasonal hidden Markov models

TL;DR

This paper introduces seasonal hidden Markov models (SHMM), proves their identifiability and MLE consistency, and demonstrates their application to precipitation data with simulated and real-world examples.

Contribution

It provides the first theoretical results on identifiability and consistency for SHMM, a class of models with time-varying transition and emission probabilities.

Findings

SHMM are identifiable under mild assumptions.

MLE is strongly consistent for SHMM.

Application to precipitation data demonstrates practical usefulness.

Abstract

In this paper, we introduce a variant of hidden Markov models in which the transition probabilities between the states, as well as the emission distributions, are not constant in time but vary in a periodic manner. This class of models, that we will call seasonal hidden Markov models (SHMM) is particularly useful in practice, as many applications involve a seasonal behaviour. However, up to now, there is no theoretical result regarding this kind of model. We show that under mild assumptions, SHMM are identifiable: we can identify the transition matrices and the emission distributions from the joint distribution of the observations on a period, up to state labelling. We also give sufficient conditions for the strong consistency of the maximum likelihood estimator (MLE). These results are applied to simulated data, using the EM algorithm to compute the MLE. Finally, we show how SHMM can be used in real world applications by applying our model to precipitation data, with mixtures of exponential distributions as emission distributions.