COMPSTAT 2022: Start Registration
View Submission - COMPSTAT2022
Title: Extending the Poisson hidden Markov model to the multilevel framework with individual random effects Authors:  Sebastian Mildiner Moraga - Utrecht University (Netherlands) [presenting]
Emmeke Aarts - Utrecht University (Netherlands)
Abstract: Hidden Markov models (HMMs) are probabilistic methods in which observations are seen as realizations of a latent Markov process with discrete states that switch over time. Moving beyond standard statistical tests, HMMs offer a statistical environment to optimally exploit the information present in multivariate time series, uncovering the latent dynamics that rule them. Although many applications of the HMM to model multivariate count data exist, so far, the support for multilevel data is restricted to non-parametric discrete random effects that apply to groups of individuals. We extend the Poisson HMM to the multilevel framework, accommodating variability between individuals with continuously distributed individual random effects, and we describe how to estimate individual and group-level parameters in a fully parametric Bayesian approach. The proposed model allows for probabilistic decoding of the sequence of hidden states based on individual-specific parameters and multivariate count time-series, and offers a framework to measure between-individual variability formally. Finally, we illustrate how to use our model to explore the latent dynamics governing complex count data with an empirical data set of multi-electrode electrophysiological measurements in macaque monkeys and a small Monte Carlo simulation.