Machine Learning for Chronology Building in Regional-Scale Synthesis

Summary

This is an abstract from the "Big Ideas to Match Our Future: Big Data and Macroarchaeology" session, at the 89th annual meeting of the Society for American Archaeology.

Chronlogical control is essential for regional-scale research in order to establish contemporaneity or temporal sequences among spatially distributed assemblages. Archaeology has benefitted from advances in radiometric dating methods, as well as statistical protocols for combining dates to achive greater precision age estimates. Yet the potential for applying these methods remains limited, requiring that samples of a narrow range of materials be recovered (often with special handling) from carefully excavated contexts. This leaves nearly all surface collections and many excavated assemblages undatable by these methods. Archaeologists also have long used artifact morphologies to estimate the age of assemblages. While widely applicable, this approach faces questions over subjectivity, replicability, and the causes of variation in artifact morphologies (and hence, their reliability for chronology). Fortunately, a suite of analytical methods, broadly termed machine learning, can combine these different dating approaches, allowing a much wider array of assemblages from many contexts--including surface assemblages–to be dated. We compare the accuracy and reliability of a suite of machine learning algorithms for chronology building, widely accessible on normal desktop computers, using assemblages from dated, excavated contexts. We then illustrate their usefulness in regional synthesis by applying them to surface assemblages extending across eastern Spain.

This Resource is Part of the Following Collections

• Society for American Archaeology 89th Annual Meeting, New Orleans, LA (2024)   •
• Big Ideas to Match Our Future: Big Data and Macroarchaeology

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 38055.0