This collection contains code for fitting common species interaction models using Hamiltonian Monte Carlo in Stan. It is intended as a learning resource, rather than flexible "production-ready" implementations. Expect that the code may contain errors and inefficiencies--use at your own risk!
Each model is contained in a sub-folder. The naming convention specifies each model as three components separated by underscores:
- the interaction type (
compet_*for competitive,conres_*for consumer-resource) - the time framework (
*_disc_*for discrete,*_cont_*for continuous) - the common name of the model in the ecological literature
Each folder has a GitHub-viewable Markdown document (.md), the RMarkdown document (.Rmd) that generated it, and a stand-alone Stan file (.stan) containing the portable model source code. R is used to either generate or import example data and interfaces Stan via the cmdstanr package. The posterior distribution of the model parameters are presented, and when possible a meaningful calculation on the posterior parameters is performed (e.g., finding the posterior distribution of the equilibrium point).
- Lotka-Volterra (discrete time)
- Lotka-Volterra (continuous time)
- Beverton-Holt
- Leslie-Gower
- Law-Watkinson
- Ricker
- Hassell
- Lotka-Volterra (discrete time)
- Lotka-Volterra (continuous time)
- Rosenzweig-MacArthur
- Beverton-Holt
- Leslie-Gower
- Nicholson-Bailey
- Arditi-Ginzburg
- Yodzis-Innes
- DeAngelis-Beddington
- Wangersky-Cunningham
Stan ecology page: additional case studies, event materials, & collected peer-reviewed papers
Spatial models for plant neighborhood dynamics in Stan: a spatially-explicit competition model wherein competitors affect focals more when they are larger and closer
We welcome contributions to this example model base. Please use model_template.Rmd in the home directory to format your contribution. Please include the following in your pull request:
- an RMarkdown source file
- a rendered Markdown file with
*_filessubfolder (e.g., containing images of plots) - a stand-alone Stan file
We also welcome bug reports & ideas to make the models more efficient.
Please note our code of conduct for contributors.