Abstract
Here, I show that powerful machine learning (ML) algorithms can efficiently classify most BSM of any given controlled experimental assemblage. In some cases, classification may reach an accuracy of 100%. No other statistical tool commonly used in taphonomy had been this successful at classification before. However, the heuristics of ML algorithms depend tightly on the objectivity in raw data collection. The use of multivariate approaches in which variables are independently scored by the analyst introduces a subjective bias. In this work, I show that different analysts producing raw data on the same testing data set can lead to widely divergent BSM classifications and interpretations using the same powerful ML algorithms. It is emphasized that until an objective non-biasing method of raw data collection is implemented, BSM classification carried out via statistical tests will remain heuristically limited. As a consequence, the application of these referents to archeological BSM does not guarantee a correct interpretation beyond the analyst expertise.
Similar content being viewed by others
References
Andrews P, Cook J (1985) Natural modifications to bones in a temperate setting. Man 20:675–691
Aramendi J, Maté-González MA, Yravedra J, Ortega MC, Arriaza MC, González-Aguilera D, Baquedano E, Domínguez-Rodrigo M, (2017) Discerning carnivore agency through the three-dimensional study of tooth pits: revisiting crocodile feeding behaviour at FLK- Zinj and FLK NN3 (Olduvai Gorge, Tanzania). Palaeogeogr Palaeoclimatol Palaeoecol (in press)
Arriaza MC, Domínguez-Rodrigo M (2016) When felids and hominins ruled at Olduvai Gorge: a machine learning analysis of the skeletal profiles of the non-anthropogenic bed I sites. Quat Sci Rev 139:43–52
Bello SM, Soligo C (2008) A new method for the quantitative analysis of cutmark micromorphology. J Archaeol Sci 35:1542–1552
Bello SM, Parfitt SA, Stringer C (2009) Quantitative micromorphological analyses of cut marks produced by ancient and modern handaxes. J Archaeol Sci 36:1869–1880
Björn-Helge M, Wehrens R (2007) The pls package: principal component and partial least squares regression in R. J Stat Softw 18:1–23
Blumenschine RJ, Marean CW, Capaldo SD (1996) Blind tests of inter-analyst correspondence and accuracy in the identification of cut marks, percussion marks, and carnivore tooth marks on bone surfaces. J Archaeol Sci 23:493–507
Boschin F, Crezzini J (2012) Morphometrical analysis on cut marks using a 3D digital microscope. Int J Osteoarchaeol 22:549–562
Dambricourt Malassé A, Moigne A-M, Singh M, Calligaro T, Karir B, Gaillard C, Kaur A, Bhardwaj V, Pal S, Abdessadok S, Chapon Sao C, Gargani J, Tudryn A, Garcia Sanz M (2016) Intentional cut marks on bovid from the Quranwala zone, 2.6 Ma, Siwalik Frontal Range, northwestern India. Comptes Rendus Palevol 15:317–339
Dimitriadou E, Hornik K, Leisch F, Meyer D, Weingessel A (2010) e1071: Misc Functions of the Department of Statistics (e1071). TU Wien (2010). https://cran.r-project.org/web/packages/e1071/e1071.pdf
Domínguez-Rodrigo M (2015) Taphonomy in early African archaeological sites: questioning some bone surface modification models for inferring fossil hominin and carnivore feeding interactions. J Afr Earth Sci 108:42–46
Domínguez-Rodrigo M, Alcalá L (2016) 3.3-million-year-old stone tools and butchery traces? More evidence needed. PaleoAnthropology 2016:46–53
Domínguez-Rodrigo M, Baquedano E (2018) Distinguishing butchery cut marks from crocodile bite marks through machine learning methods. Nat Sci Rep 8. https://doi.org/10.1038/s41598-018-24071-1
Domínguez-Rodrigo M, Barba R (2005) A study of cut marks on small-sized carcasses and its application to the study of cut-marked bones from small mammals at the FLK Zinj site. J Taphon 3:121–134
Domínguez-Rodrigo M, Pickering TR (2010) A mutivariate approach for discriminating bone accumulations created by spotted hyenas and leopards: harnessing actualistic data from east and southern Africa. J Taphon 8:155–179
Domínguez-Rodrigo M, Pickering TR (2016) The meat of the matter: an evolutionary perspective on human carnivory. Azania: Archaeological Research in Africa 0:1–29
Domínguez-Rodrigo M, Barba R, Egeland CP (2007) Deconstructing Olduvai: a taphonomic study of the bed I sites. Springer Science & Business Media, Berlin
Domínguez-Rodrigo M, de Juana S, Galán AB, Rodríguez M (2009) A new protocol to differentiate trampling marks from butchery cut marks. J Archaeol Sci 36:2643–2654
Domínguez-Rodrigo M, Pickering TR, Bunn HT (2010) Configurational approach to identifying the earliest hominin butchers. Proc Natl Acad Sci 107:20929–20934
Domínguez-Rodrigo M, Pickering TR, Bunn HT (2011) Reply to McPherron et al.: doubting Dikika is about data, not paradigms. Proc Natl Acad Sci 108:E117–E117
Domínguez-Rodrigo M, Bunn HT, Yravedra J (2014) A critical re-evaluation of bone surface modification models for inferring fossil hominin and carnivore interactions through a multivariate approach: application to the FLK Zinj archaeofaunal assemblage (Olduvai Gorge, Tanzania). Quat. Int. 322–323:32–43
Domínguez-Rodrigo M, Saladié P, Cáceres I, Huguet R, Yravedra J, Rodríguez-Hidalgo A, Martín P, Pineda A, Marín J, Gené C, Aramendi J, Cobo-Sánchez L (2017) Use and abuse of cut mark analyses: the Rorschach effect. J Archaeol Sci 86:14–23
Domínguez-Rodrigo M, Wonmin B, Arampatzis G, Baquedano E, Yravedra J, Maté-González MA, Koumoutsakos P (2018) Automated identification and deep classification of cut marks on bones and its paleoanthropological implications (submitted)
Egeland CP, Domínguez-Rodrigo M (2008) Taphonomic perspectives on hominid site use and foraging strategies during bed II times at Olduvai Gorge, Tanzania. J Hum Evol 55:1031–1052
Fariña RA, Tambusso PS, Varela L, Czerwonogora A, Di Giacomo M, Musso M, Bracco R, Gascue A (2014) Arroyo del Vizcaíno, Uruguay: a fossil-rich 30-ka-old megafaunal locality with cut-marked bones. Proc Bio Sci Royal Soc 281:20132211
Fernandez-Jalvo Y, Andrews P (2016) Atlas of taphonomic identifications: 1001+ images of fossil and recent mammal bone modification. Springer, Berlin
Ferraro JV, Plummer TW, Pobiner BL, Oliver JS, Bishop LC, Braun DR, Ditchfield PW, Seaman JW 3rd, Binetti KM, Seaman JW Jr, Hertel F, Potts R (2013) Earliest archaeological evidence of persistent hominin carnivory. PLoS One 8:e62174
Galán AB, Rodríguez M, de Juana S, Domínguez-Rodrigo M (2009) A new experimental study on percussion marks and notches and their bearing on the interpretation of hammerstone-broken faunal assemblages. J Archaeol Sci 36:776–784
Harris JA, Marean CW, Ogle K, Thompson J (2017) The trajectory of bone surface modification studies in paleoanthropology and a new Bayesian solution to the identification controversy. J Hum Evol 110:69–81
Hastie T, Tibshirani R, Friedman J (2016) The elements of statistical learning. Springer, New York
Holen SR, Deméré TA, Fisher DC, Fullagar R, Paces JB, Jefferson GT, Beeton JM, Cerutti RA, Rountrey AN, Vescera L, Holen KA (2017) A 130,000-year-old archaeological site in southern California, USA. Nature 544:479–483
Hu L, Huang MW, Ke S-W, Tsai CF 2016(2016) The distance function effect on k-nearest neighbour classification for medical datasets. Springerplus 5(1):1304. https://doi.org/10.1186/s40064-016-2941-7
James EC, Thompson JC (2015) On bad terms: problems and solutions within zooarchaeological bone surface modification studies. Environ Archaeol 20:89–103
de Juana S, Galán AB, Domínguez-Rodrigo M (2010) Taphonomic identification of cut marks made with lithic handaxes: an experimental study. J Archaeol Sci 37:1841–1850
Kuhn M (2017) C5.0 decision trees and rule-based models. https://cran.r-project.org/web/packages/C50/C50.pdf
Kuhn M, Johnson K (2013) Applied predictive modeling. Springer, New York
Lantz B (2013) Machine learning with R. Packt Publishing Ltd., Birmingham
Lyman R (1994) Vertebrate taphonomy. Cambridge University Press, Cambridge
Liaw A, Wiener M (2002) Classification and regression by randomForest. R News 2(3):18–22
Lyman RL (1987) Archaeofaunas and butchery studies: a taphonomic perspective. Adv Archaeo Meth Theory 10:249–337
Maté González MÁ, Yravedra J, González-Aguilera D, Palomeque-González JF, Domínguez-Rodrigo M (2015) Micro-photogrammetric characterization of cut marks on bones. J Archaeol Sci 62:128–142
Maté-González MÁ, Palomeque-González JF, Yravedra J, González-Aguilera D, Domínguez-Rodrigo M (2016) Micro-photogrammetric and morphometric differentiation of cut marks on bones using metal knives, quartzite, and flint flakes. Archaeol Anthropol Sci 1–12
McPherron SP, Alemseged Z, Marean CW, Wynn JG, Reed D, Geraads D, Bobe R, Béarat HA (2010) Evidence for stone-tool-assisted consumption of animal tissues before 3.39 million years ago at Dikika, Ethiopia. Nature 466:857–860
Merritt SR (2012) Factors affecting Early Stone Age cut mark cross-sectional size: implications from actualistic butchery trials. J Archaeol Sci 39:2984–2994
Meyer D, Leisch F, Hornik K (2003) The support vector machine under test. Neurocomputing 55(1–2):169–186
Moretti E, Arrighi S, Boschin F, Crezzini J, Aureli D, Ronchitelli A (2015) Using 3D microscopy to analyze experimental cut marks on animal bones produced with different stone tools. Ethnobio Lett 6:267–275
Njau J (2012) Paleontology. Reading pliocene bones. Science 336:46–47
Organista E, Domínguez-Rodrigo M, Yravedra J, Uribelarrea D, Arriaza MC, Ortega MC, Mabulla A, Gidna A, Baquedano E (2017) Biotic and abiotic processes affecting the formation of BK Level 4c (Bed II, Olduvai Gorge) and their bearing on hominin behavior at the site. Palaeogeogr Palaeoclimatol Palaeoecol
Pante MC, Blumenschine RJ, Capaldo SD, Scott RS (2012) Validation of bone surface modification models for inferring fossil hominin and carnivore feeding interactions, with reapplication to FLK 22, Olduvai Gorge, Tanzania. J Hum Evol 63:395–407
Pante MC, Muttart MV, Keevil TL, Blumenschine RJ, Njau JK, Merritt SR (2017) 1. A new high-resolution 3-D quantitative method for identifying bone surface modifications with implications for the Early Stone Age archaeological record. J Hum Evol 102:1–11
Venables WN, Ripley BD (2002) Modern applied statistics with S, 4th edn. Springer, New York
Weihs C, Ligges U, Luebke K, Raabe N (2005) klaR analyzing German business cycles. In: Baier D, Decker R, Schmidt-Thieme L (eds) Data analysis and decision support. Springer-Verlag, Berlin, pp 335–343
Wolpert DH (1996) The existence of a priori distinctions between learning algorithms. Neural Comput 8:1391–1420
Yravedra J, Maté-González MÁ, Palomeque-González JF, Aramendi J, Estaca-Gómez V, San Juan Blazquez M, García Vargas E, Organista E, González-Aguilera D, Arriaza MC et al (2017) A new approach to raw material use in the exploitation of animal carcasses at BK (Upper Bed II, Olduvai Gorge, Tanzania): a micro-photogrammetric and geometric morphometric analysis of fossil cut marks. Boreas 46:860–873
Acknowledgments
MDR thanks D. Lieberman and the Human Evolutionary Biology Department at Harvard and the Royal Complutense College at Harvard, where this research was conducted. I thank Ruth Blasco, Nick Conard, and two anonymous reviewers for their constructive comments on an earlier draft of this manuscript. I also thank my colleagues P. Saladié, I. Cáceres, R. Huguet, J. Yravedra, A. Rodríguez-Hidalgo, P. Martín, A. Pineda, J. Marín, C. Gené, J. Aramendi, and L. Cobo-Sánchez for their joint work in the experimental analysis that led to inter-analyst comparisons of BSM.
Funding
This work was carried out with support from a Research Salvador Madariaga grant to MDR (Ministry of Education, Culture and Sport, Spain, Ref. PRX16/00010).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Domínguez-Rodrigo, M. Successful classification of experimental bone surface modifications (BSM) through machine learning algorithms: a solution to the controversial use of BSM in paleoanthropology?. Archaeol Anthropol Sci 11, 2711–2725 (2019). https://doi.org/10.1007/s12520-018-0684-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12520-018-0684-9