Typicality and predictive distributions in discriminant function analysis

Lyle W Konigsberg, Susan R. Frankenberg

Research output: Contribution to journalArticle

Abstract

While discriminant function analysis is an inherently Bayesian method, researchers attempting to estimate ancestry in human skeletal samples often follow discriminant function analysis with the calculation of frequentist-based typicalities for assigning group membership. Such an approach is problematic because it fails to account for admixture and for variation in why individuals may be classifijied as outliers or nonmembers of particular groups. This article presents an argument and methodology for employing a fully Bayesian approach in discriminant function analysis applied to cases of ancestry estimation. The approach requires adding the calculation, or estimation, of predictive distributions as the fijinal step in ancestry-focused discriminant analyses. The methods for a fully Bayesian multivariate discriminant analysis are illustrated using craniometrics from identifijied population samples within the Howells published data. The article also presents ways to visualize predictive distributions calculated in more than three dimensions, explains the limitations of typicality measures, and suggests an analytical route for future studies of ancestry and admixture based in discriminant function analysis.

Original languageEnglish (US)
Pages (from-to)31-44
Number of pages14
JournalHuman biology
Volume90
Issue number1
DOIs
StatePublished - Dec 1 2018

Fingerprint

Discriminant Analysis
ancestry
discriminant analysis
Bayesian theory
Bayes Theorem
outlier
researchers
methodology
distribution
analysis
Multivariate Analysis
sampling
Research Personnel
Population
method
calculation

Keywords

  • Admixture
  • Bayesian analysis
  • Outliers
  • Posterior probability
  • Tukey depth

ASJC Scopus subject areas

  • Ecology, Evolution, Behavior and Systematics
  • Genetics
  • Genetics(clinical)

Cite this

Typicality and predictive distributions in discriminant function analysis. / Konigsberg, Lyle W; Frankenberg, Susan R.

In: Human biology, Vol. 90, No. 1, 01.12.2018, p. 31-44.

Research output: Contribution to journalArticle

@article{88eb14f998c446f8bd12d70c4541491a,
title = "Typicality and predictive distributions in discriminant function analysis",
abstract = "While discriminant function analysis is an inherently Bayesian method, researchers attempting to estimate ancestry in human skeletal samples often follow discriminant function analysis with the calculation of frequentist-based typicalities for assigning group membership. Such an approach is problematic because it fails to account for admixture and for variation in why individuals may be classifijied as outliers or nonmembers of particular groups. This article presents an argument and methodology for employing a fully Bayesian approach in discriminant function analysis applied to cases of ancestry estimation. The approach requires adding the calculation, or estimation, of predictive distributions as the fijinal step in ancestry-focused discriminant analyses. The methods for a fully Bayesian multivariate discriminant analysis are illustrated using craniometrics from identifijied population samples within the Howells published data. The article also presents ways to visualize predictive distributions calculated in more than three dimensions, explains the limitations of typicality measures, and suggests an analytical route for future studies of ancestry and admixture based in discriminant function analysis.",
keywords = "Admixture, Bayesian analysis, Outliers, Posterior probability, Tukey depth",
author = "Konigsberg, {Lyle W} and Frankenberg, {Susan R.}",
year = "2018",
month = "12",
day = "1",
doi = "10.13110/humanbiology.90.1.01",
language = "English (US)",
volume = "90",
pages = "31--44",
journal = "Human Biology",
issn = "0018-7143",
publisher = "Wayne State University Press",
number = "1",

}

TY - JOUR

T1 - Typicality and predictive distributions in discriminant function analysis

AU - Konigsberg, Lyle W

AU - Frankenberg, Susan R.

PY - 2018/12/1

Y1 - 2018/12/1

N2 - While discriminant function analysis is an inherently Bayesian method, researchers attempting to estimate ancestry in human skeletal samples often follow discriminant function analysis with the calculation of frequentist-based typicalities for assigning group membership. Such an approach is problematic because it fails to account for admixture and for variation in why individuals may be classifijied as outliers or nonmembers of particular groups. This article presents an argument and methodology for employing a fully Bayesian approach in discriminant function analysis applied to cases of ancestry estimation. The approach requires adding the calculation, or estimation, of predictive distributions as the fijinal step in ancestry-focused discriminant analyses. The methods for a fully Bayesian multivariate discriminant analysis are illustrated using craniometrics from identifijied population samples within the Howells published data. The article also presents ways to visualize predictive distributions calculated in more than three dimensions, explains the limitations of typicality measures, and suggests an analytical route for future studies of ancestry and admixture based in discriminant function analysis.

AB - While discriminant function analysis is an inherently Bayesian method, researchers attempting to estimate ancestry in human skeletal samples often follow discriminant function analysis with the calculation of frequentist-based typicalities for assigning group membership. Such an approach is problematic because it fails to account for admixture and for variation in why individuals may be classifijied as outliers or nonmembers of particular groups. This article presents an argument and methodology for employing a fully Bayesian approach in discriminant function analysis applied to cases of ancestry estimation. The approach requires adding the calculation, or estimation, of predictive distributions as the fijinal step in ancestry-focused discriminant analyses. The methods for a fully Bayesian multivariate discriminant analysis are illustrated using craniometrics from identifijied population samples within the Howells published data. The article also presents ways to visualize predictive distributions calculated in more than three dimensions, explains the limitations of typicality measures, and suggests an analytical route for future studies of ancestry and admixture based in discriminant function analysis.

KW - Admixture

KW - Bayesian analysis

KW - Outliers

KW - Posterior probability

KW - Tukey depth

UR - http://www.scopus.com/inward/record.url?scp=85055902694&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055902694&partnerID=8YFLogxK

U2 - 10.13110/humanbiology.90.1.01

DO - 10.13110/humanbiology.90.1.01

M3 - Article

VL - 90

SP - 31

EP - 44

JO - Human Biology

JF - Human Biology

SN - 0018-7143

IS - 1

ER -