DARth ABC: Analysis of the Dated Archaeological Record through Approximate Bayesian Computation

This repository contains a set of scripts in R that implement the method described in "Analysis of the Abundance of Radiocarbon Samples without the Sum of Probability Distributions" by Miguel de Navascués, Concetta Burgarella and Mattias Jakobsson (2024) doi:10.5281/zenodo.13381596.

Vignette

A vignette is included with a simple example. After reading the publication, use this as a starting point for undersanding how to apply the functions in your own analysis.

Scripts

Main file: DARthABC.R

File that contains the functions needed for the analysis of the abundance of radiocarbon data.

Files to reproduce analysis presented in Navascués et al. (2024) doi:10.5281/zenodo.13381596

• MethodEvaluation.R analyses a toy model with different models (probability distribution model vs. Poisson (data as counts) model) and summary statistics (based on SPD vs. based on CRA dates).
• MethodEvaluationPlots.R plots the results from the analysis performed in MethodEvaluation.R.
• Bevan_all.R analyses the data from Bevan et al. (2017) doi:10.1073/pnas.1709190114. Requires to download the data set published in Bevan (2017) doi:10.14324/000.ds.10025178.
• Bevan_cereals.R analyses a subset of those data that correponds to samples identified as barley or wheat.

Other scrpits

Other files correspond to work in progress and will be described in the future.

Requirements (R packages)

These scripts have been tested in R 4.1.2. The following packages are required (with corresponding version tested):

DARthABC:

• rcarbon (1.5.1)
• weights (4.1.2)
• extraDistr (1.9.1)

Scripts that apply the method to data:

• abcrf (1.9)
• doSNOW (1.0.20)
• doParallel (1.0.17)
• doRNG (1.8.6)