Automatic Data Quality Assessment of Hydrographic Surveys Taking Into Account Experts' Preferences
Résumé
Data quality assessment of hydrographic surveys is a complex problem, since context dependent acquisition conditions using multiple sensors contribute to numerous data imperfections, which have different consequences depending on the intended final product. In this work we propose a generic methodology integrating experts' preferences through multi-criteria preference models with data quality techniques to generate explainable overall quality assessments of hydrographic surveys which depend on the expected end-uses. Four hydrographic surveys of different geographic locations and contrasting characteristics were studied, according to the preferences of an acoustician, an oceanographer, and a hydrographer. The obtained results indicate that each survey was appropriately evaluated, indicating the reasons that led to the specific assessment. Index Terms-hydrographic survey, data quality assessment, preference modelling, multi-criteria decision aiding.
Origine : Fichiers produits par l'(les) auteur(s)