Abstract: "The deep complexity of uncertain data modelling has resisted to [sic] general solutions so far. Instead, a diversity of modelling approaches has been proposed over the years, but few systems actually have been built. The DUCK calculus is one recent ambitious rule- based attempt to mo...

Ausführliche Beschreibung

Bibliographische Detailangaben

Hauptverfasser:	Kießling, Werner (VerfasserIn), Köstler, Gerhard (VerfasserIn), Güntzer, Ulrich (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	München 1992
Schriftenreihe:	Technische Universität <München>: TUM 9237
Schlagworte:	Künstliche Intelligenz Artificial intelligence Deductive databases
Zusammenfassung:	Abstract: "The deep complexity of uncertain data modelling has resisted to [sic] general solutions so far. Instead, a diversity of modelling approaches has been proposed over the years, but few systems actually have been built. The DUCK calculus is one recent ambitious rule- based attempt to model uncertainty on the grounds of established probability theory as typically used e.g. in medical diagnosis. This paper describes how deductive database technology can be exploited for prototyping of a system for uncertain reasoning. In particular we discuss the issues of ADT-ideas in Datalog by using interpreted predicates. Moreover we show that for safety reasons logic programming and current Datalog optimizers must be upgraded to deal with semantic optimization in form of subsumption New differential least fixpoint operators, customized for subsumption optimization, are provided. Finally, we outline the design and implementation of DUCK-Demonstrator/1.1 which serves as a research vehicle for ongoing studies of uncertain reasoning phenomena and for optimization of vague queries.
Beschreibung:	32 S.