Approximating data in constraint databases

Abstract

Approximate representation of any spatio-temporal variable, by some interpolation function, is necessary when it is measured only sporadically. This paper argues that the approximate representation can be captured by a constraint database. Since constraint databases can be queried via standard query languages - such as relational algebra, SQL and Datalog - this provides an immediate benefit for flexible querying of the data. We propose a concrete system that implements a version of this approach. We also add beyond the standard queries new ones like cartogram similarity queries and an advanced graphical user interface with 3-D animation of GIS-based data. © 2000 Springer-Verlag Berlin Heidelberg.