Table 1. Model selection results for the two regression fitting models using all radiocarbon dates and only human dates.
| Model | R2 | logLik | AIC | Δ | w | |
|---|---|---|---|---|---|---|
| All | Exponential | 0.7765 | 12715.43 | -25424.87 | 8101.36 | 0.00 |
| dates | Logistic | 0.9241 | 16766.12 | -33526.23 | 0.00 | 1.00 |
| Human | Exponential | 0.8304 | 30017.73 | -60029.47 | 4168.31 | 0.00 |
| dates | Logistic | 0.9027 | 32101.89 | -64197.78 | 0.00 | 1.00 |
R2 is the mean square degree of fit for the SPDs, logLik is the log-likelihood of the regression models, AIC is the calculated Akaike’s Information Criteria, Δ is the AIC difference and w is Akaike’s weight, or model probability that is calculated as exp (-0.5 * ΔAIC of each model) divided by the sum of these values for all models.