Bioarchaeological research in northwestern Argentina: analyses by classification and regression trees
DOI:
https://doi.org/10.6092/issn.1973-2201/1168Abstract
This bioarchaeological study investigates biological characteristic of human groups localised in archaeological sites in the department Belén in Northwestern Argentina. They cover a period from the Formative period till the Incas and the Spanish Conquest. To understand the variability over the time as well as the systems of interchange and the social differentiation data from 35 human skulls were analysed. The 49 discrete variables per skull represent facial features and morphological attributes. The aim was to establish relationships between individuals from the different sites or to find attributes which can discriminate between the archaeological sites. One specific problem in analysing these data was the significant percentage of missing values since some of the skulls were not complete. The method used for this analysis was CART (Classification and Regression Trees). CART can handle cases with incomplete data by identifying surrogate predictor variables. The result show a classification tree of five terminal nodes with the discriminating factors "hueso epiptérico", "foramen oval incomplete", "sutura infraorbitaria complete" and "hueso astérionico". These characteristics are known as heritabilities and do not show an ecological influence. This allow us to propose that there has been a considerable social and spatial mobility over the time in the observer area, with a longer preservation of the genotype in the communities of earlier origin and an increasing heterogeneity towards the later periods.How to Cite
Wiese, B., Colaneri, M. G., Bru de Labanda, E., & Raya, R. (2001). Bioarchaeological research in northwestern Argentina: analyses by classification and regression trees. Statistica, 61(1), 135–142. https://doi.org/10.6092/issn.1973-2201/1168
Issue
Section
Articles
License
Copyright (c) 2001 Statistica
This journal is licensed under a Creative Commons Attribution 3.0 Unported License (full legal code).
Authors accept to transfer their copyrights to the journal.
See also our Open Access Policy.