Uncertainty as a Function of Expertise

Abstract

Information Source Tracking (IST) method is an elegant framework proposed for the representation and manipulation of inaccurate data. In IST, sources contribute data in the database and the reliability of data depends on the reliability of the contributing sources. While IST is capable of managing inaccurate data, it is incapable of accommodating reliability revision of sources and apparently can not assign reliability to sources as a function of their expertise. In this paper we introduce the concept of context in the IST to accommodate the notion of context dependent reliability of sources. We show that the introduction of the concept of contexts greatly enhances the naturality and the conceptual modeling capability of the basic IST model. We also show that under the assumption of reliability independence among sources in different contexts, an extended IST database is IST reducible. Finally, we extend the definitions of the relational operators proposed for basic IST to accommodate contexts, and show that the extended operators are information preserving.

This research was supported in part by grants from the Natural Sciences and Engineering Research Council of Canada and the Fonds Pour Formation De Chercheurs Et L’Aide À La Recherche of Quebec.

This author’s research was additionally supported in part by grants from the Canadian Commonwealth Scholarship and Fellowship Plan, and the University of Dhaka, Bangladesh. The author is on leave from the University of Dhaka, Bangladesh.