Evaluating machine learning techniques for archaeological lithic sourcing: a case study of flint in Britain

Tom Elliot,Robert Morse,Duane Smythe,Ashley Norris

doi:10.1038/s41598-021-87834-3

Tom Elliot, Robert Morse + Show 2 more

Open Access

https://doi.org/10.1038/s41598-021-87834-3

Copy DOI

Abstract

It is 50 years since Sieveking et al. published their pioneering research in Nature on the geochemical analysis of artefacts from Neolithic flint mines in southern Britain. In the decades since, geochemical techniques to source stone artefacts have flourished globally, with a renaissance in recent years from new instrumentation, data analysis, and machine learning techniques. Despite the interest over these latter approaches, there has been variation in the quality with which these methods have been applied. Using the case study of flint artefacts and geological samples from England, we present a robust and objective evaluation of three popular techniques, Random Forest, K-Nearest-Neighbour, and Support Vector Machines, and present a pipeline for their appropriate use. When evaluated correctly, the results establish high model classification performance, with Random Forest leading with an average accuracy of 85% (measured through F1 Scores), and with Support Vector Machines following closely. The methodology developed in this paper demonstrates the potential to significantly improve on previous approaches, particularly in removing bias, and providing greater means of evaluation than previously utilised.

Highlights

It is 50 years since Sieveking et al published their pioneering research in Nature on the geochemical analysis of artefacts from Neolithic flint mines in southern Britain
The first step to evaluate the machine learning techniques was first to interrogate the structure of the geochemical data using t-distributed Stochastic Neighbour Embedding (t-SNE)
The results show the viability of using machine learning techniques to classify flint in Britain at scale, with Random Forest showing the greatest overall potential, followed closely by Support Vector Machines

Summary

Introduction

With the advent of Machine Learning techniques, such separation is routinely possible, using iterative methodologies that improve on their results through validation of reliable training data The utility of such approaches has been seen more widely in Archaeology, including towards remote sensing and prediction or classification of archaeological sites[11,12,13], the recording and creation of artefact typologies[14,15,16,17,18,19], and more recently for lithic sourcing 20–24. These include basic prerequisites such as inadequate sampling from individual geological sources for machine learning techniques to effectively learn from[23], to perhaps more

Objectives

Methods

Results

Discussion

Conclusion