Hidden Markov models for animal movement and other ecological data
Sunday, 26 June 2022 (Full day)
Presenters: Roland Langrock, Sina Mews, Théo Michelot, Richard Glennie and Timo Adam
Hidden Markov models (HMMs) are flexible statistical models for sequences of observations that are driven by underlying states. Over the last decade, HMMs have become increasingly popular within the ecological community as general-purpose tools for the analysis of animal movement and general behaviour data, but also as a modelling framework within which many capture-recapture and occupancy models can be embedded to facilitate statistical inference.
This 1-day workshop will introduce the HMM framework, comprising a mix of theoretical lectures and hands-on practical components using R. It will be expected that participants are familiar with basic statistics and probability theory (e.g. what is a probability distribution? what is conditional probability? what is maximum likelihood?). In the theoretical sessions, the following topics will be covered:
- overview & basic model formulation
- fitting an HMM to data
- model selection & model checking
- state decoding
- incorporating covariates, random effects and seasonality
- extensions of the basic model formulation (e.g. multivariate time series)
These techniques will be illustrated primarily using movement data, but are applicable also to other ecological time series data (e.g. capture-recapture). The practical components will focus on the use of the R packages moveHMM and momentuHMM. Basic knowledge of the free software R would be advantageous, but is not required.
Cost: R1,000 per person in person (US$70), R650 per person virtual (US$45)