Abstract

Consistency measures provide an indication on how much a dataset satisfies a set of integrity constraints, which is useful for comparing, integrating and cleaning datasets. This work presents the notion of consistency measures and provides an evaluation of the cognitive adequacy of these measures. It evaluates the impact on the consistency measures of different parameters (overlapping size, external distance, internal distance, crossing length, and touching length) and the relative size of geometries involved in a conflict. While a human-subject testing supports our hypotheses with respect to the parameters, it rejects the significance of the relative size of geometries as a component of the consistency measures. © 2011 Springer-Verlag.