• 14C Calibration with JAGS


    In this post I would like to continue what I have published already 4 years ago. This time I want to show how to calibrate a ¹⁴C-date using well-known Bayesian tools, specifically JAGS. JAGS stands for Just Another Gibbs Sampler, and is a program for analysing ‘Bayesian hierarchical models using Markov Chain Monte Carlo (MCMC) simulation’. This makes it clearly overqualified for our simple task, but it nicely shows a different perspective on the problem and also makes obvious the possibilities of how to extend this task even further.

  • 14C Calibration the Bayesian way


    With what follows I want to demonstrate that the basic functionality of Oxcal is reflected in this tiny bit of code:

    likelihood <- function(proposal){dnorm(calf(proposal),bp,std)}
    collector <- bp <- 3600; std <- 30; last_prob <- likelihood(bp)
    for (i in 1:10000) {
      proposal <- rnorm(1, tail(collector,1), 3*std)
      proposal_prob <- likelihood(proposal)
      if ((proposal_prob/last_prob) > runif(1))
        last_prob <- proposal_prob
      } else {
        proposal <- tail(collector,1)
      collector <- c(collector, proposal)

  • 14C Calibration 1

    Calibration of 14C dates is one of the more common tasks an archaeologists has to do. I will not go into the details of the details why calibration is necessary here (maybe in another post), but I want to give in this post and in some follow ups hands on instruction how a calibration can be archieved using R. This series will be published also at the ISAAKiel-Repository.

  • Blog Entry Number One

    To get my new website finally online, I want to add something to the blog section. Here in the future mostly stuff should be published that is related to archaeology and statistics using R.

    To start I want to add a bit that helps to understand the relationship of several items within a collection of data. To be more specific, how can I explore the correlation of lets say pottery types or in this case animal species in the archaeological remains of different sites. Do specific combination occure regularily, so that it might be the case that they are functional related?

subscribe via RSS