Analysis of a Bayesian Network Methodology for Site Similarity Assessment

Author(s): Deborah Leishman; Jean Pike

Year: 2023


This is an abstract from the "The Expanding Bayesian Revolution in Archaeology" session, at the 88th annual meeting of the Society for American Archaeology.

We present work on a methodology that sits at the intersection of architecture, archaeology, and Bayesian statistics to expand the quantity of architectural data considered in analysis of precontact architectures. Two sites are examined as possible precedents for Pueblo Bonito at Chaco Canyon, NM: the late ninth-century McPhee Pueblo in Colorado and Gila Butte–Snaketown Phase large Hohokam ballcourt at Snaketown, Arizona. Using architectural analysis and Bayesian networks, the degree of each site’s similarity to Pueblo Bonito ca. AD 860 is represented as a probability of similarity. Results show both sites have a high similarity to Pueblo Bonito’s Western Arc. We focus here on Bayesian Networks as the core of the methodology. These Networks are based on Bayes Theorem, utilize graphical representations of relationships between the multiple architectural variables, and specify a sound propagation method to determine the probability of similarity. Previous use of Bayesian statistics in archaeology mostly supports chronological dating. Bayesian Networks go beyond this and are necessary due to the large number of variables which are both quantitative (ratio of height to width) and qualitative (winter solstice alignment). Our findings indicate the methodology supports a sound probabilistic assessment of similarity that can be easily replicated.

Cite this Record

Analysis of a Bayesian Network Methodology for Site Similarity Assessment. Deborah Leishman, Jean Pike. Presented at The 88th Annual Meeting of the Society for American Archaeology. 2023 ( tDAR id: 474283)

Spatial Coverage

min long: -124.365; min lat: 25.958 ; max long: -93.428; max lat: 41.902 ;

Record Identifiers

Abstract Id(s): 36913.0