Improving RPNI algorithm using minimal message length

Authors:
Petr Hoffmann
Affiliations:
Department of Software and Computer Science Education, Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic
Venue:
AIAP'07 Proceedings of the 25th conference on Proceedings of the 25th IASTED International Multi-Conference: artificial intelligence and applications
Year:
2007

Citing 6
Cited 0

Genetic Algorithms in Search, Optimization and Machine Learning

Genetic Algorithms in Search, Optimization and Machine Learning
Inference of regular languages using model simplicity

ACSC '01 Proceedings of the 24th Australasian conference on Computer science
Results of the Abbadingo One DFA Learning Competition and a New Evidence-Driven State Merging Algorithm

ICGI '98 Proceedings of the 4th International Colloquium on Grammatical Inference
Regular Grammatical Inference from Positive and Negative Samples by Genetic Search: the GIG Method

ICGI '94 Proceedings of the Second International Colloquium on Grammatical Inference and Applications
What Is the Search Space of the Regular Inference?

ICGI '94 Proceedings of the Second International Colloquium on Grammatical Inference and Applications
Learning DFA from Simple Examples

ALT '97 Proceedings of the 8th International Conference on Algorithmic Learning Theory

Quantified Score

Hi-index	0.00

Visualization

Abstract

There are known several state merging methods for regular language inference from positive and negative samples. One of them is the RPNI algorithm that merges pairs of states of the prefix tree acceptor of the positive samples in a fixed order assuring consistency of the resulting automaton. The resulting automaton need not be the optimal one. By prohibiting some merges done by the original RPNI algorithm it is possible to get a better automaton. We propose a new method of searching pairs of states which should not be merged using genetic algorithms and a random walk. The improvement over the original RPNI algorithm is evaluated experimentally.