A New Bayesian Approach for Estimating Chronological Events and Phases with ChronoModel

Summary

This is an abstract from the "Constructing Chronologies I: Stratification and Correlation" session, at the 86th annual meeting of the Society for American Archaeology.

Many issues in archaeology concern the issue of phasing—the beginning, end, and duration of a given period. We define a “Phase” as a group of Events (Event dates) that share common features. Currently used Phase models implemented in many software packages employ statistical models that concentrate posterior Event dates. Alternatively, we introduce a new statistical definition for “Phase,” which does not require additional parameters (Phase boundaries) and is based directly on the Event model—a new, robust Bayesian statistical model combining contemporaneous dates, stratigraphic constraints, and the temporal ordering of phases that will act on Event dates. A posteriori, the beginning, end, and duration of a Phase are estimated from posterior Event dates using MCMC techniques. Moreover, we introduce a new means for characterizing the temporal characteristics of Phases via calculations of their time range, gap range, and transition range at a given confidence level. Furthermore, this framework allows one to calculate the cumulative frequency of archaeological events (e.g., 14C measurements) within a given Phase, and these tempo plot trajectory curves illustrate the mean number of dated events prior to a given date within a Phase. The Bayesian statistical methods described above are implemented in the ChronoModel software package.

This Resource is Part of the Following Collections

• Society for American Archaeology 86th Annual Meeting, Online (2021)   •
• Constructing Chronologies I: Stratification and Correlation

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 32904