Is learning RFSAs better than learning DFAs?

  • Authors:

  • Pedro García;José Ruiz;Antonio Cano;Gloria Alvarez

  • Affiliations:

  • Universidad Politécnica de Valencia, Valencia, Spain;Universidad Politécnica de Valencia, Valencia, Spain;Universidad Politécnica de Valencia, Valencia, Spain;Pontificia Universidad Javeriana - Seccional Cali., Cali, Colombia

  • Venue:
  • CIAA'05 Proceedings of the 10th international conference on Implementation and Application of Automata
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Inference of RFSAs has been recently presented [1] as an alternative to inference of DFAs if the target language has been obtained by a random generation of NFAs. We propose in this paper the algorithm RPNI2, which is a variation of the previous RPNI, that also outputs DFAs as hypothesis. The experiments done using the same data as in [1] show that RPNI2 has an error rate very similar to the rate obtained in the inference of RFSAs, but the size of the hypothesis is substantially smaller.