Louhari Nourine
Abstract:
Over the last decades, many contributions have been done in the field of database dependencies. From well-known dependencies for database design (functional, inclusion, multi-valued, \ldots dependencies) to dependencies for data quality (conditional functional dependencies, matching dependencies, denial dependencies, \ldots), there is a plethora of such dependencies.
Understanding the syntax and the semantics of each dependency is sometimes a bit difficult. For instance, CFD are defined with the notion of pattern tuples and therefore requires some effort to understand something relatively easy.
In this talk, we point out an ongoing work for defining a general approach to deal with data dependencies over inconsistant and/or approximate data.
The main motivation of this work comes data in Astrophysics, where data values are mainly given as intervals of real numbers.