Learning regular languages using RFSAs
Theoretical Computer Science - Special issue: Algorithmic learning theory
Learning Regular Languages Using Nondeterministic Finite Automata
CIAA '08 Proceedings of the 13th international conference on Implementation and Applications of Automata
A randomised inference algorithm for regular tree languages
Natural Language Engineering
Hi-index | 0.00 |
We present in this paper a new family of algorithms for regular languages inference from complete presentation. Every algorithm of this family, on input of the sets of words (D+,D−), obtains for every x in D+ at least a non deterministic finite automaton (NFA) which accepts x and is consistent with D−. This automaton is, besides, irreducible in the sense that any further merging of states accepts words of D−. The output of the algorithm is a NFA which consists of the collection of NFAs associated to each word of D+. Every algorithm of the family converges to a automaton for the target language. We also present the experiments done to compare one of the algorithms of the family with two other well known algorithms for the same task. The results obtained by our algorithm are better, both in error rate as in the size of the output.