Repairs and consistent answers for inconsistent probabilistic spatio-temporal databases

We formally introduce the concept of repair and consistent answer for inconsistent probabilistic spatio-temporal databases. We start by defining the syntax and semantics of SPOT databases, a declarative framework that has been explored in recent years for the representation of spatio-temporal data with uncertainty expressed as probability intervals. In this framework we study two types of repairs, that is, minimal modifications that lead to consistent databases: maximal consistent subsets and probability interval expansion. We also extend the concept of consistent answer to this framework and find that this can be done in several different ways. In emphasizing tractable cases we propose polynomial-time algorithms for computing consistent answers and repairs based on probability interval expansion, and experimentally validate our approach.