Hidden Markov Mixture Autoregressive Models: Parameter Estimation

TL;DR

This paper proposes a new mixture autoregressive model with hidden Markov structure, and adapts EM and Baum-Welch algorithms for efficient parameter estimation.

Contribution

It introduces a parsimonious mixture autoregressive model with hidden Markov weights and develops modified algorithms for parameter estimation.

Findings

Effective parameter estimation via modified EM and Baum-Welch algorithms

Model captures complex dependencies in time series data

Provides a new framework for mixture autoregressive modeling

Abstract

This report introduces a parsimonious structure for mixture of autoregressive models, where the weighting coefficients are determined through latent random variables as functions of all past observations. These variables follow a hidden Markov model. We modify EM and Baum-Welch algorithms to estimate the parameters of the model.