Distinguishing Discoid and Centripetal Levallois methods through machine learning

Abstract

In this paper, we apply Machine Learning (ML) algorithms to study the differences between Discoid and Centripetal Levallois methods. For this purpose, we have used experimentally knapped flint flakes, measuring several parameters that have been analyzed by seven ML algorithms. From these analyses, it has been possible to demonstrate the existence of statistically significant differences between Discoid products and Centripetal Levallois products, thus contributing with new data and a new method to this traditional debate. The new approach enabled differentiating the blanks created by both knapping methods with an accuracy >80% using only ten typometric variables. The most relevant variables were maximum length, width to the 25%, 50% and 75% of the flake length, external and internal platform angles, maximum width and number of dorsal scars. This study also demonstrates the advantages of the application of multivariate ML methods to lithic studies.

Osteometric Study of Metapodial Bones and Phalanges as Indicators of the Behavioural Ecology of Modern Reindeer (<i>Rangifer tarandus</i>) and Implications for Reconstruction of Paleo Mobility

“…Thus, Kappa values of 0.3-0.6 show reasonable agreement. Values higher than these indicate a high agreement between the expected accuracy and the documented one (Domínguez-Rodrigo & Baquedano, 2018;González-Molina et. al 2020).…”

“…Lastly, we applied Machine Learning (ML) techniques, testing five models to identify the best possible classification algorithms for our purposes. Machine learning techniques have started to be applied in the field of archaeozoology and it has been demonstrated they can represent powerful tools for classification purposes (Lefebvre et al 2016;Cifuentes-Alcobendas et al, 2019;González-Molina et. al 2020;.…”

“…The remaining 30% (40% for metacarpals, metatarsals, and anterior proximal phalanges) of the sample was retained as the test sample. This avoided dealing with the bias/variance trade-off (González-Molina et. al 2020;.…”

“…After an ensemble of trees is generated, the model uses a vote to combine the trees predictions. This method avoids overfitting through the use of OOB ("out of the bag") technique, estimating how many iterations are needed to minimize the OOB error (Lantz, 2013;Kuhn & Johnson, 2013;Domínguez-Rodrigo & Baquedano, 2018;González-Molina et. al 2020).…”

Artifact3-D: New software for accurate, objective and efficient 3D analysis and documentation of archaeological artifacts

“…[ 148 , 149 ]), their proliferation has been slow within archaeology owing to the fragmentary and varied nature of the archaeological record. While there have been some interesting recent attempts at applying machine learning to the study of lithics [ 150 – 152 ], hominins [ 153 – 155 ], fauna [ 156 ], pottery [ 157 ], art [ 158 , 159 ], built structures [ 160 ], geoarchaeology [ 161 – 164 ], and remote sensing [ 165 ], our sample sizes and training sets are often too small for widespread use of these methods. Moreover, many of these studies require drastic reductions in resolution while also remaining susceptible to variability in orientation and inter-user variability.…”