Statistical tools to directly model underlying population dynamics using date datasets (radiocarbon and other).
Various model structures can be compared in a robust formal model comparison framework. Continuous Piecewise Linear (CPL) models are infinitely flexible, and can recover complex population dynamics if enough data is available. Other simpler models include: Uniform, Exponential, Gaussian, Cauchy, Sinusoidal, Logistic and Power law. Taphonomic loss included optionally as a power function.
Posterior parameter estimates of population models, using model likelihoods and a weak uniform prior.
Package also calibrates 14C samples, generates Summed Probability Distributions (SPD), and performs SPD simulation analysis to generate a Goodness-of-fit test for the best selected model. Continuous Piecewise Linear (CPL) models that are flexible to estimate any complex population dynamics
Install from CRAN, then load
install.packages('ADMUR')
library('ADMUR')Refer to the vignette 'guide' for detailed support and examples.
vignette('guide', package = 'ADMUR')Please contact [email protected] in the first instance to make suggestions, report bugs or request help.
This package accompanies the following paper:
Timpson A., Barberena R., Thomas M. G., Mendez C., Manning K. 2020. "Directly modelling population dynamics in the South American Arid Diagonal using 14C dates",Philosophical Transactions B. https://doi.org/10.1098/rstb.2019.0723.
Citations available as follows:
citation(package='ADMUR')ADMUR was written in collaboration with:
- University College London
- Department of Genetics, Evolution and Environment
- Molecular and Cultural Evolution Laboratory
- UCL Research Software Development
- Max Planck Institute for the Science of Human History
- Department of Archaeology
- Pan African Evolution ResearchGroup
Special thanks to Yoan Diekmann for his influential inferential input.
Also thanks to the following who have reported bugs, requested additional functionality, offered constructive criticism, or provided other advice:
- Gregor Seyer
- Uwe Ligges
- Prof Brian Ripley
- Enrico Crema
- Ricardo Fernandes
- Mark G. Thomas
- Robert Staniuk