Ripple Down Rules
Up
PAC-Bayesian
Bayesian Transduction
Bayes Point Machines
Adatpive Margin Machines
Sparsity
Ordinal Regression
Proximity Learning
Performance Assessment
Concept Learning
Ripple Down Rules
Algorithmic Luckiness
Semidefinite Programming
Informative Vector Machines
Learning to Fight
ROC Curve Bounds
Poisson Networks
Approximate Bayesian Inference
Drivatars

 

 

Algebra and machine induction of ripple down rules

A ripple down rule is a list of rules, each of which may be connected to another ripple down rule, specifying exceptions, hence, rules can be patched locally. This makes them an interesting representation scheme for knowledge acquisition, since an expert may insert new rules, which only have effects in the given context of the parent rule. A ripple down rule algebra provides the basis for transformations of rules (like reduction, transformation into flat lists of rules, ...), a knowledge revision algorithm provides automatic induction and revision of RDR knowledge bases.

References

Up | PAC-Bayesian | Bayesian Transduction | Bayes Point Machines | Adatpive Margin Machines | Sparsity | Ordinal Regression | Proximity Learning | Performance Assessment | Concept Learning | Ripple Down Rules | Algorithmic Luckiness | Semidefinite Programming | Informative Vector Machines | Learning to Fight | ROC Curve Bounds | Poisson Networks | Approximate Bayesian Inference | Drivatars

This site was last updated 29-10-2004