Figure 6.
Tuned MCMC results for the two-pool model with a parameters on the linear-scale. The left-most column shows an MCMC trace from a single chain. Kernel density estimates of the marginal posterior for each parameter and bivariate scatter plots are produced using every 300th sample from four independent MCMC chains, after burn-in. The autocorrelation function (ACF) for a single chain is shown in (c), indicating that every 300th sample is approximately independent.