A Model for Multigranular Data and Its Integrity
Volume 28, Issue 1 (2017), pp. 45–78
Pub. online: 1 January 2017
Type: Research Article
Received
1 December 2016
1 December 2016
Accepted
1 February 2017
1 February 2017
Published
1 January 2017
1 January 2017
Abstract
Data involving spatial and/or temporal attributes are often represented at different levels of granularity in different source schemata. In this work, a model of such multigranular data is developed, which supports not only the usual order structure on granules, but also lattice-like join and disjointness operators for relating such granules in much more complex ways. In addition, a model for multigranular thematic attributes, to which aggregation operators are applied, is provided. Finally, the notion of a thematic multigranular comparison dependency, generalizing ordinary functional and order dependencies but specifically designed to model the kinds of functional and order dependencies which arise in the multigranular context, and in particular incorporating aggregation into the definition of the constraint, is developed.