Learning context-free grammars from partially structured examples

Yasubumi Sakakibara, Hidenori Muramatsu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

33 Citations (Scopus)


In this paper, we consider the problem of inductively learning context-free grammars from partially structured examples. A structured example is represented by a string with some parentheses inserted to indicate the shape of the derivation tree of a grammar. We show that the partially structured examples contribute to improving the efficiency of the learning algorithm. We employ the GA-based learning algorithm for context-free grammars using tabular representations which Sakakibara and Kondo have proposed previously [7], and present an algorithm to eliminate unnecessary nonterminals and production rules using the partially structured examples at the initial stage of the GA-based learning algorithm. We also show that our learning algorithm from partially structured examples can identify a context-free grammar having the in- tended structure and is more flexible and applicable than the learning methods from completely structured examples [5].

Original languageEnglish
Title of host publicationGrammatical Inference
Subtitle of host publicationAlgorithms and Applications - 5th International Colloquium, ICGI 2000, Proceedings
EditorsArlindo L. Oliveira
PublisherSpringer Verlag
Number of pages12
ISBN (Print)9783540452577
Publication statusPublished - 2000
Externally publishedYes
Event5th International Colloquium on Grammatical Inference, ICGI 2000 - Lisbon, Portugal
Duration: 2000 Sept 112000 Sept 13

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other5th International Colloquium on Grammatical Inference, ICGI 2000

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Learning context-free grammars from partially structured examples'. Together they form a unique fingerprint.

Cite this