A Novel Ensemble Machine Learning Approach for Bioarchaeological Sex Prediction

Evan Muzzall

doi:10.3390/technologies9020023

Abstract

I present a novel machine learning approach to predict sex in the bioarchaeological record. Eighteen cranial interlandmark distances and five maxillary dental metric distances were recorded from n = 420 human skeletons from the necropolises at Alfedena (600–400 BCE) and Campovalano (750–200 BCE and 9–11th Centuries CE) in central Italy. A generalized low rank model (GLRM) was used to impute missing data and Area under the Curve—Receiver Operating Characteristic (AUC-ROC) with 20-fold stratified cross-validation was used to evaluate predictive performance of eight machine learning algorithms on different subsets of the data. Additional perspectives such as this one show strong potential for sex prediction in bioarchaeological and forensic anthropological contexts. Furthermore, GLRMs have the potential to handle missing data in ways previously unexplored in the discipline. Although results of this study look promising (highest AUC-ROC = 0.9722 for predicting binary male/female sex), the main limitation is that the sexes of the individuals included were not known but were estimated using standard macroscopic bioarchaeological methods. However, future research should apply this machine learning approach to known-sex reference samples in order to better understand its value, along with the more general contributions that machine learning can make to the reconstruction of past human lifeways.

Highlights

Accurate sex prediction of archaeological skeletal remains is a fundamental step for reconstructing biological and demographic profiles of past humans
The SuperLearner algorithm has the highest Area Under the Curve—Receiver Operating Characteristic (AUC-ROC) for all six bony regions while ranger is a close second for the face, vault, base, cranial, and combined craniodental data
An important potential contribution of this research is that it reframes the problem of sex estimation as a predictive one and does not rely on assumptions of p-values, traditional hypothesis testing, or causal inference approaches

Summary

Introduction

Accurate sex prediction of archaeological skeletal remains is a fundamental step for reconstructing biological and demographic profiles of past humans. After an archaeological site is surveyed and excavated and unknown human remains are identified, documented, and recovered, the sex and age of deceased individuals are commonly estimated using macroscopic methods of the pelvis, skull, and teeth [1,2,3]. The pelvis and cranium might provide conflicting sex estimation results even within the same individual. This process is further complicated by other aspects, of age, as tooth crown calcification and eruption and bone epiphyseal fusion are useful until early adulthood when 3rd molars erupt and bony ossification centers fuse skeletal elements into their final, united shapes. Cranial suture, and sternal rib end methods are used to predict age in individuals through later stages of adulthood, albeit with wider margins of error

Methods

Results

Conclusion