Probabilities and predicates in knowledge refinement

Authors:
Roy Rada
Affiliations:
National Library of Medicine, Bethesda, MD
Venue:
PKWBS-W'84 Proceedings of the 1984 IEEE conference on Principles of knowledge-based systems
Year:
1984

Citing 4
Cited 0

Data Structures and Algorithms

Data Structures and Algorithms
Genetic algorithms as function optimizers

Genetic algorithms as function optimizers
A domain-independent production-rule system for consultation programs

IJCAI'79 Proceedings of the 6th international joint conference on Artificial intelligence - Volume 2
Characterizing search spaces

IJCAI'83 Proceedings of the Eighth international joint conference on Artificial intelligence - Volume 2

Quantified Score

Hi-index	0.00

Visualization

Abstract

To facilitate refinement a probabilistic, rule-based expert system should be designed so that small changes in the structure correspond to small changes in the function or performance. Numbers (reflecting probabilities) have an ordering that makes their manipulation more straightforward than the manipulation of predicates. A simple credit assignment and probability adjustment strategy for improving numbers in a probabilistic, rule-based expert system has been developed. Re-arranging predicates in rules is a more difficult problem that requires a relationship or ordering be first established among predicates.