Decision Support with Imprecise Data for Consumers

  • Authors:

  • Gergely Lukács

  • Affiliations:

  • -

  • Venue:
  • Soft-Ware 2002 Proceedings of the First International Conference on Computing in an Imperfect World
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Only imperfect data is available in many decision situations, which therefore plays a key role in the decision theory of economic science. It is also of key interest in computer science, among others when integrating autonomous information systems: the information in one system is often imperfect from the view of another system. The case study for the present work combines the two issues: the goal of the information integration is to provide decision support for consumers, the public. By the integration of an electronic timetable for public transport with a geographically referenced database, for example, with rental apartments, it is possible to choose alternatives, for example, rental apartments from the database that have a good transport connection to a given location. However, if the geographic references in the database are not sufficiently detailed, the quality of the public transport connections can only be characterized imprecisely. This work focuses on two issues: the representation of imprecise data and the sort operation for imprecise data. The proposed representation combines intervals and imprecise probabilities. When the imprecise data is only used for decision making with the Bernoulli-principle, a more compact representation is possible without restricting the expressive power. The key operation for decision making, the sorting of imprecise data, is discussed in detail. The new sort operation is based on so called 驴-cuts, and is particularly suitable for consumer decision support.