In Defense of Intervals
Abstract
Why intervals or upper and lower probabilities? And how should such a system work? I claim that if a system for the representation of uncertainty is to be useful, it must be possible to obtain from it constraints on decisions in the face of uncertainty. By this, I mean that, as in real life , even if we are in a state of some uncertainty, there should ordinarily be some actions that are ruled out by rational considerations. Interval representations of uncertainty will accomplish this.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1988
- Accession Number
- ADA250539
Entities
People
- Henry Kyburg Jr.
Organizations
- University of Rochester