EAA2020: Abstract
Abstract is part of session #464:
Title & Content
Title:
Reality is what I recognise?
Choice of factors and ML in social archaeology
Choice of factors and ML in social archaeology
Content:
Especially in prehistoric archaeology datasets are often large, fuzzy and important factors can hardly be determined. When using a model as heuristic process one is often required to select factors that seem relevant to the hypothesis. Predominantly this choice is based on a social model fed with empircal data, essentially blending quantitative with theoretical archaeology.
However, it becomes challenging when the latter is strongly biased or there is simply not enough information to justify a prior selection of relevant indicators. Furthermore, if exploring the relams of quantiative and basic explanatory statistics, the data is often not grouped strongly enough or exhibit clear patterns. Machine Learning can, albeit it's own theoretical and methodological problems, distinguish patterns ins fuzzy and incomplete datasets. Therefore it offers a new perspective, allowing one to integrate more data, without prior selection and implementing social theory at a later stage of hypothesis building, hopefully generate a more objective model outcome and therefore a better understanding of the past.
This talk aims to give a theoretical perspective on the power of ML in archaeological social theory.
However, it becomes challenging when the latter is strongly biased or there is simply not enough information to justify a prior selection of relevant indicators. Furthermore, if exploring the relams of quantiative and basic explanatory statistics, the data is often not grouped strongly enough or exhibit clear patterns. Machine Learning can, albeit it's own theoretical and methodological problems, distinguish patterns ins fuzzy and incomplete datasets. Therefore it offers a new perspective, allowing one to integrate more data, without prior selection and implementing social theory at a later stage of hypothesis building, hopefully generate a more objective model outcome and therefore a better understanding of the past.
This talk aims to give a theoretical perspective on the power of ML in archaeological social theory.
Keywords:
social theory, statistics, fuzzy datasets
Downloads:
authors
Main authors:
Chiara G. M. Girotto1
Co-author:
Affiliations:
1 Independent Researcher