Testing archaeological predictive models: a rough guide

Verhagen, Philip

In: Posluschny, A. ; Lambers, K. ; Herzog, I. (Hrsgg.): Layers of Perception. Proceedings of the 35th International Conference on Computer Applications and Quantitative Methods in Archaeology (CAA), Berlin, 2.-6. April 2007. Koll. Vor- u. Frühgesch. 10. Bonn 2008, pp. 285-291

[thumbnail of 09_01_verhagen_predictive.pdf]

PDF, English
Download (365kB) | Terms of use

For citations of this document, please do not use the address displayed in the URL prompt of the browser. Instead, please cite with one of the following:

Abstract

Archaeological predictive modelling is an essential instrument for archaeological heritage management in the Netherlands. It is used to decide where to do archaeological survey in the case of development plans. However, very little attention is paid to testing the predictions made. Model quality is established by means of peer review, rather than by quantitative criteria. In this paper the main issues involved with predictive model testing are discussed. The potential of resampling methods for improved predictive model quality is investigated, and the problems associated with obtaining representative test data sets are highlighted.

Document type:	Book Section
Version:	Secondary publication
Date Deposited:	30 Mar 2010 11:09
Faculties / Institutes:	Research Project, Working Group > Individuals
DDC-classification:	Alte Geschichte, Vor- und Frühgeschichte, Archäologie
Controlled Keywords:	Archäologie, Informatik, Computerunterstütztes Verfahren, Computervisualistik, Interdisziplinäre Forschung, Naturwissenschaften, Methode
Subject (classification):	History of the ancient world to ca. 499
Countries/Regions:	Europe north and west of Italy (Antiquity)