Dataset Information

Advancing predictive modeling in archaeology: An evaluation of regression and machine learning methods on the Grand Staircase-Escalante National Monument.

ABSTRACT: Predictive models are central to both archaeological research and cultural resource management. Yet, archaeological applications of predictive models are often insufficient due to small training data sets, inadequate statistical techniques, and a lack of theoretical insight to explain the responses of past land use to predictor variables. Here we address these critiques and evaluate the predictive power of four statistical approaches widely used in ecological modeling-generalized linear models, generalized additive models, maximum entropy, and random forests-to predict the locations of Formative Period (2100-650 BP) archaeological sites in the Grand Staircase-Escalante National Monument. We assess each modeling approach using a threshold-independent measure, the area under the curve (AUC), and threshold-dependent measures, like the true skill statistic. We find that the majority of the modeling approaches struggle with archaeological datasets due to the frequent lack of true-absence locations, which violates model assumptions of generalized linear models, generalized additive models, and random forests, as well as measures of their predictive power (AUC). Maximum entropy is the only method tested here which is capable of utilizing pseudo-absence points (inferred absence data based on known presence data) and controlling for a non-representative sampling of the landscape, thus making maximum entropy the best modeling approach for common archaeological data when the goal is prediction. Regression-based approaches may be more applicable when prediction is not the goal, given their grounding in well-established statistical theory. Random forests, while the most powerful, is not applicable to archaeological data except in the rare case where true-absence data exist. Our results have significant implications for the application of predictive models by archaeologists for research and conservation purposes and highlight the importance of understanding model assumptions.

SUBMITTER: Yaworsky PM

PROVIDER: S-EPMC7529236 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Advancing predictive modeling in archaeology: An evaluation of regression and machine learning methods on the Grand Staircase-Escalante National Monument.

Yaworsky Peter M PM Vernon Kenneth B KB Spangler Jerry D JD Brewer Simon C SC Codding Brian F BF

PloS one 20201001 10

Predictive models are central to both archaeological research and cultural resource management. Yet, archaeological applications of predictive models are often insufficient due to small training data sets, inadequate statistical techniques, and a lack of theoretical insight to explain the responses of past land use to predictor variables. Here we address these critiques and evaluate the predictive power of four statistical approaches widely used in ecological modeling-generalized linear models, ...[more]

PMID: 33002016

Similar Datasets

Project description:Grand Staircase-Escalante National Monument is a federally protected area found in central southern Utah. Designated in 1996 by President William J. Clinton, it was recently reduced in size by President Donald J. Trump in a proclamation that turned the one large monument into three smaller ones. A long-term, standardized study of the bees had been conducted from 2000-2003, revealing 660 species. The bee communities of the area are characterized by being spatially heterogeneous; most of the bees occur in isolated areas, with only a few being both abundant and widespread. Here we examine what affect the recent resizing of the monument has on the number, and ecology, of the bees now excluded from monument boundaries. Using the new monument boundaries and the geographic coordinates associated with each bee, we derived new species lists for each of the three monuments, and compared them to each other, and to the excluded lands. All three monuments now protect unique faunas, with Bray-Curtis similarity values not exceeding 0.59%. Each monument now harbors species not found in the other two monuments. We found that 84 bee species are no longer protected by any of the three monuments. These 84 species were not concentrated in one area that is now excluded, but were scattered throughout the newly excluded lands. For some of the excluded bee species, there is no evidence that they are rare or imperiled, being widespread throughout the west. However, there is a concentration of bees in the southern and eastern former monument lands that represent range extensions from nearby hot deserts. In addition to numerous range extensions, the list of excluded bees also contains several undescribed species (newly discovered species that have not yet been named and described by taxonomists) and morphospecies (individuals that are morphologically distinct, but that require additional research before species designations can be made). This indicates that the bee communities housed in these excluded areas would benefit from additional scientific inquiry. The areas now excluded from monument protections house a greater proportion of the original GSENM bee community than any of the three new monument units. We conclude this paper by discussing what the smaller monuments might mean for bee conservation in this hot spot of bee biodiversity and suggest that bee communities here and elsewhere should be taken into account when conservation decisions are being made.

Dataset Information

Advancing predictive modeling in archaeology: An evaluation of regression and machine learning methods on the Grand Staircase-Escalante National Monument.

Publications

Advancing predictive modeling in archaeology: An evaluation of regression and machine learning methods on the Grand Staircase-Escalante National Monument.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets