A Test-Case Based Approach to Bayesian Knowledge Base Incompleteness Detection and Correction
Abstract
This work develops tools and techniques to identify particular Bayesian Knowledge Base (BKB) incompletenesses, and to modify the existing knowledge-base (KB) structure to correct these problems. The methodology performs manually or automatically, informing the user of either problems causing the incompleteness, or of details resulting from the automatic knowledge-base correction. The proposed methodology is designed for integration with BVAL, to augment BVAL's validation techniques.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1996
- Accession Number
- ADA320697
Entities
People
- Louise J. Lyle
Organizations
- Air Force Institute of Technology