Title: A Bayesian network to represent a data quality model

Authors: Angelica Caro, Coral Calero, Houari A. Sahraoui, Mario Piattini

Addresses: Department of Computer Science and Information Technologies, University of Bio Bio, Chillan, Chile. ' Information Systems and Technologies Department, Alarcos Research Group, UCLM-INDRA Research and Development Institute, University of Castilla-La Mancha, Paseo de la Universidad, 4, Ciudad Real, Spain. ' Department d'Informatique et de Recherche Operationnelle, Universite de Montreal, CP 6128 succ, Centre Ville, Montreal, QC H3C 3J7, Canada. ' Information Systems and Technologies Department, Alarcos Research Group, UCLM-INDRA Research and Development Institute, University of Castilla-La Mancha, Paseo de la Universidad, 4, Ciudad Real, Spain

Abstract: Web portals provide data to many people where data consumers need to assess Data Quality (DQ). In our previous work a Portal Data Quality Model (PDQM) was developed. PDQM is focused on data consumers| perspective and is composed by 33 attributes appropriate for DQ evaluation. Now, we have organised these attributes into a generic and operational structure. Considering the uncertainty inherent in perception of quality, we decided to use a probabilistic approach, using Bayesian Networks (BNs). This paper, explains the definition of the BN structure that supports PDQM.

Keywords: data quality; information quality; web portals; Bayesian networks; data quality models; data quality evaluation.

DOI: 10.1504/IJIQ.2007.016392

International Journal of Information Quality, 2007 Vol.1 No.3, pp.272 - 294

Published online: 26 Dec 2007 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article