Estimation of age at death from the pubic symphysis and the auricular surface of the ilium using a smoothing procedure

Abstract

We discuss here the estimation of age at death from two indicators (pubic symphysis and the sacro-pelvic surface of the ilium) based on four different osteological series from Portugal, GreatBritain, South Africa or USA (European origin). These samples and the scoring system of the two indicators were used by Schmitt et al. (2002), applying the methodology proposed by Lucy et al. (1996). In the present work, the same data was processed using a modification of the empirical method proposed by Lucy et al. (2002). The various probability distributions are estimated from training data by using kernel density procedures and Jackknife methodology. Bayes's theorem is then used to produce the posterior distribution from which point and interval estimates may be made. This statistical approach reduces the bias of the estimates to less than 70% of what was obtained by the initial method. This reduction going up to 52% if knowledge of sex of the individual is available, and produces an age for all the individuals that improves age at death assessment. Estimation of age at death from the pubic symphysis and the auricular surface of the ilium using a smoothing procedure AbstractWe discuss here the estimation of age at death from two indicators (pubic symphysis and the sacro-pelvic surface of the ilium) based on four different osteological series from Portugal, Great-Britain, South Africa or USA (European origin). These samples and the scoring system of the two indicators were used by Schmitt et al. (2002), applying the methodology proposed by Lucy et al. (1996). In the present work, the same data was processed using a modification of the empirical method proposed by Lucy et al. (2002). The various probability distributions are estimated from training data by using kernel density procedures and Jackknife methodology. Bayes's theorem is then used to produce the posterior distribution from which point and interval estimates may be made. This statistical approach reduces the bias of the estimates to less than 70% of what was obtained by the initial method. This reduction going up to 52% if knowledge of sex of the individual is available, and produces an age for all the individuals that improves age at death assessment.

New approach to age estimation of male and female adult skeletons based on the morphological characteristics of the acetabulum

“…In this way, it is indispensable to improve aging techniques, thereby reducing the error and generating population standards based on specific and complete documented samples. Bayesian inference presents new opportunities for improving the accuracy of age estimates and appears to provide robust and appropriate results [16,17,38,[67][68][69].…”

Quantitative Analysis of the Morphological Changes of the Pubic Symphyseal Face and the Auricular Surface and Implications for Age at Death Estimation

“…Many different methodological and statistical approaches have been used to improve the accuracy of age estimation or to combine different existing methods: artificial neural networks ; nature‐inspired data mining methods ; smoothing procedure ; and transition analysis . The new approach we explored utilized the 3D digital nature of CT and laser scans and a single, simple way to mathematically describe the surface changes, namely the extrinsic curvature.…”

Optimal trait scoring for age estimation

“…Three classes of nonparametric approaches have so far been used in the literature. These include categorical models that Konigsberg and Frankenberg () refer to as “contingency table paleodemography,” kernel‐smoothing (Lucy et al, ; Love and Müller, ; Lucy et al, ; Müller et al, ; Roksandic and Armstrong, ; Martins et al, ), and the Sugeno fuzzy integral (Anderson et al, ). It is relatively rare to find discussion of why a nonparametric method was selected over a parametric approach.…”

“…In contrast, following Müller and coworkers (Love and Müller, ; Müller et al, ), some authors (Coqueugniot et al, ; Roksandic and Armstrong, ) have argued for using nonparametric models—specifically kernel smoothing—to characterize age progression in skeletal traits, although this approach also was given in earlier literature on dental traits (Lucy et al, ; Aykroyd et al, ; Lucy et al, ). Martins et al () use kernel densities for the pubic symphysis and auricular surface citing the earlier work by Lucy et al ().…”