Toward Efficient Bayesian Approaches to Inference in Hierarchical Hidden Markov Models for Inferring Animal Behavior

Giada Sacchi,Ben Swallow

doi:10.3389/fevo.2021.623731

Abstract

The study of animal behavioral states inferred through hidden Markov models and similar state switching models has seen a significant increase in popularity in recent years. The ability to account for varying levels of behavioral scale has become possible through hierarchical hidden Markov models, but additional levels lead to higher complexity and increased correlation between model components. Maximum likelihood approaches to inference using the EM algorithm and direct optimization of likelihoods are more frequently used, with Bayesian approaches being less favored due to computational demands. Given these demands, it is vital that efficient estimation algorithms are developed when Bayesian methods are preferred. We study the use of various approaches to improve convergence times and mixing in Markov chain Monte Carlo methods applied to hierarchical hidden Markov models, including parallel tempering as an inference facilitation mechanism. The method shows promise for analysing complex stochastic models with high levels of correlation between components, but our results show that it requires careful tuning in order to maximize that potential.

Highlights

In a crucial moment for the environment, while climate is rapidly changing and putting flora and fauna under great threat, it is important to investigate the evolution of ecological populations, in order to estimate demographic parameters and forecast their future developments.The study of individuals of the same species and their behaviors, how they constitute the populations in which they exist and how such populations evolve is called population ecology (King et al, 2010)
Behavioral models can be employed in population ecology to reflect the patterns in animal behaviors and movements, enabling the analysis and identification of different modes over time
A basic Hidden Markov Model is a doubly stochastic process with observable state-dependent processes {Yt}Tt=1 controlled by duenpdeenrdlyeinntgdissttartiebuptriooncess{sfei }sNi={S1 t(}LTt=eo1,s through the so-called stateBarajas and Michelot, 2018)

Summary

Introduction

In a crucial moment for the environment, while climate is rapidly changing and putting flora and fauna under great threat, it is important to investigate the evolution of ecological populations, in order to estimate demographic parameters and forecast their future developments.The study of individuals of the same species and their behaviors, how they constitute the populations in which they exist and how such populations evolve is called population ecology (King et al, 2010). Hidden Markov Models (HMMs) and associated state-switching models are becoming increasingly common time series models in ecology, since they can be used to model animal movement data and infer various aspects of animal behavior (Leos Barajas et al, 2017). They are able to model the propensity to persist in such behaviors over time and to explain the serial dependence typically found, by enabling the connection of observed data points to different underlying ecological processes and behavioral modes

Objectives

Methods

Results

Conclusion