COMBINING MEASUREMENTS OR DATES USING HIERARCHICAL BAYESIAN STATISTICS: THE EVENT MODEL - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2015

COMBINING MEASUREMENTS OR DATES USING HIERARCHICAL BAYESIAN STATISTICS: THE EVENT MODEL

Anne Philippe

Résumé

The most widely used methods for measurements or dates combination are the weighted mean and the central age model which are based on maximum likelihood (ML) estimation. However, these techniques do not take into account errors occurring between measurements or dates and the ML estimation does not operate when considering individual errors. This article proposes a new approach for combining measurements or dates through the Event model which is based on Bayesian statistics. This approach allows the estimation of a mean (mean of measurements or date of an Event from different datings) by introducing individual errors on each observation in addition to the experimental errors made during laboratory experiments. In particular, Event model implies that dates are coming from assumed contemporaneous artefacts. A prior information on these individual errors is introduced using the shrinkage density. This modelling provides a very simple way to automatically penalize outlying data with minimal assumptions about Bayesian parameters. The mathematical formulation is explained in details and many application examples are exhibited in different archaeological situations involving radiocarbon, lumi-nescence and archaeomagnetic results. This new combination procedure is also applied to the wiggle-matching process in dendrochronological dating. Calculations are based on MCMC numerical techniques and can be performed using ChronoModel software which is freeware and multiplatform.
Fichier principal
Vignette du fichier
Event.pdf (1.88 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-01162404 , version 1 (10-06-2015)
hal-01162404 , version 2 (05-08-2015)
hal-01162404 , version 3 (10-12-2015)

Identifiants

  • HAL Id : hal-01162404 , version 1

Citer

Philippe Lanos, Anne Philippe. COMBINING MEASUREMENTS OR DATES USING HIERARCHICAL BAYESIAN STATISTICS: THE EVENT MODEL. 2015. ⟨hal-01162404v1⟩
435 Consultations
1393 Téléchargements

Partager

Gmail Facebook X LinkedIn More