Elamur Research⌕
☰
home / papers / 2011 / Developing a New Approach for Arabic Morphological Analysis and Generation

Developing a New Approach for Arabic Morphological Analysis and Generation

Paper · 2011 · Mourad Gridach and Noureddine Chenfour

Summary

The paper describes AMAUT (Arabic Morphological Automaton), a system for both analysing and generating Arabic words. Arabic builds most nouns and verbs by placing root consonants (for example "ktb") into vowel-and-consonant templates called schemes, and then attaching prefixes and suffixes. The authors encode these roots, schemes and affixes in a morphological database written in their own XML-based language, XMODEL, and drive analysis with automata built on this structure. The system is implemented in Java with XML, which the authors say makes it portable. For each word it returns the affixes, stem, scheme and morpho-syntactic features. They compared it with the Xerox Arabic analyzer and Otakar Smrz's ElixirFM on 975 words from ALECSO texts. AMAUT left the fewest words unanalysed, and the authors report roughly 90% average performance. Its weaknesses are that it needs diacritized input, has no English glosses, and accepts non-existent words. The work matters because a compact, template-driven lexicon can reduce lexicon size and simplify maintenance for a highly derivational language.

Key points

  • AMAUT models Arabic morphemes in three groups: roots (mostly 3–4 letters), schemes (templates with slots for root letters) and affixes. The morphological database is written in XMODEL.
  • The test set had 975 ALECSO words: 481 nouns, 362 verbs and 132 particles.
  • Words left unanalysed on this set:
    • Xerox: 84 (39 nouns, 16 verbs, 29 particles)
    • ElixirFM: 152 (76 nouns, 21 verbs, 55 particles)
    • AMAUT: 21
  • For context, the paper describes the competing resources this way:
    • Buckwalter: about 77,800 stems covering 45,000 lexical items
    • Xerox: 4,930 roots and 400 patterns, yielding about 90,000 stems
  • Known limitations of AMAUT:
    • It cannot process undiacritized text.
    • It provides no English glossary.
    • It overgenerates. Figure 15 shows invalid words accepted as real ones, and the planned fix is to filter output against an existing lexicon such as Buckwalter's or Xerox's.

Interesting ideas

  • I1. Store derivational templates plus roots instead of listing every stem, so one rule covers many words and the lexicon stays small and easy to update.
  • I2. One shared knowledge base serves both analysis and generation, so the two directions cannot drift apart.
  • I3. The analyzer's output shows its derivation (affixes, stem, scheme, features) rather than only a tag.
  • I4. Generative rules overgenerate, so the authors propose pruning their output against an attested lexicon from other systems.
  • I5. The comparison counts unanalysed words per part of speech on one shared sample, and Figure 15 shows the system accepting nonsense words.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (tokens): English derivational families (-ness, -able, -ize, un-) could be stored as explicit templates over base lemmas in crate en, shrinking the lexicon while keeping each analysis traceable to a rule.
  • I2 (back): the generator could produce inflected forms from lemma plus UD features using the same tables the analyzer reads, guaranteeing that whatever is generated can be parsed back.
  • I3 (tokens): each lemma/feature decision could carry its segmentation and the template that licensed it, so later stages can explain why a word received its analysis.
  • I4 (learn): a rule-generated analysis could be accepted into silver data only when a corpus or another witness also attests the form, an agreement check that filters out invented words.
  • I5 (learn): morphology evaluation could report coverage broken down by POS and add a negative control of fabricated words, where any confident analysis counts as an error.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 2Research subjects and languages the work is about.Semitic languages · 2,908 worksMorphology · 405 works

Cite

Cite as Mourad Gridach and Noureddine Chenfour. 2011. Developing a New Approach for Arabic Morphological Analysis and Generation. arXiv:1101.5494.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.1101.5494 · arXiv 1101.5494

@misc{gridach2011developing,
  title = {{Developing a New Approach for Arabic Morphological Analysis and Generation}},
  author = {Mourad Gridach and Noureddine Chenfour},
  year = {2011},
  doi = {10.48550/arxiv.1101.5494},
  eprint = {1101.5494},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.1101.5494}
}

References

24 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Al-Sughaiyer, I. A., and Al-Kharashi, I. A. (2004), ‘Arabic morphological analysis techniques: A comprehensive survey’, Journal of the American Society for Information Science and Technology, 55(3), 189–213. ↗
  2. Altantawy M., Habash N., Rambow O., and Ibrahim S. (2010), ‘Morphological Analysis and Generation of Arabic Nouns: A Morphemic Functional Approach’, In Proceedings of the Language Resource and Evaluation Conference, Malta.
  3. Attia M. (2005), ‘Developing a Robust Arabic Morphological Transducer Using Finite State Technology’, 8th Annual CLUK Research Colloquium. Manchester, UK. find
  4. Attia, M. (2006), ‘An Ambiguity-Controlled Morphological Analyzer for Modern Standard Arabic Modelling Finite State Networks’, The Challenge of Arabic for NLP/MT Conference, the British Computer Society, London. find
  5. Atwell E., Al-Sulaiti L., Al-Osaimi S., Abu Shawar B. (2004), ‘A Review of Arabic Corpus Analysis Tools’, JEP-TALN 04, Arabic Language Processing, Fès, 19-22 April 2004. find
  6. Beesley KR. (1996), ‘Arabic Finite-State Morphological Analysis and Generation’, Proceedings of the 16th conference on Computational linguistics, Vol 1. Copenhagen, Denmark: Association for Computational Linguistics, pp 89-94.
  7. Beesley KR. (2000), ‘Finite-State Non-Concatenative Morphotactics’, Proceedings of the Fifth Workshop of the ACL Special Interest Group in Computational Phonology, p. 1-12, August 6, 2000, Luxembourg. ↗
  8. Buckwalter T. (2002). ‘Buckwalter Arabic Morphological Analyzer Version 1.0’. Linguistic Data Consortium, University of Pennsylvania, LDC Catalog No.: LDC2002L49. find
  9. Buckwalter T. (2004). ‘Buckwalter Arabic morphological analyzer version 2.0’. find
  10. Cavalli-Sforza V., Soudi A. and Teruko M. (2000), ‘Arabic Morphology Generation Using a Concatenative Strategy’, in proceedings of the 1st Conference of the North American Chapter of the Association for Computational Linguistics, Seattle, USA.
  11. Darwish K. (2002), ‘Building a Shallow Morphological Analyzer in One Day’, proceedings of the workshop on Computational Approaches to Semitic Languages in the 40th Annual Meeting of the Association for Computational Linguistics (ACL-02). Philadelphia, PA, USA. find
  12. De Roeck, A. N., and Al-Fares W. (2000), ‘A Morphologically Sensitive Clustering Algorithm for Identifying Arabic Roots’, in the 38th Annual Meeting of the ACL. Hong Kong.
  13. Diab M., Hacioglu K. and Jurafsky D. (2007a), ‘Automated Methods for Processing Arabic Text: From Tokenization to Base Phrase Chunking’, in Arabic Computational Morphology: Knowledge based and Empirical Methods, eds A. van den Bosch and A. Soudi. find
  14. El-Sadany, T. A., and Hashish, M. A. (1989), ‘An Arabic Morphological System’, IBM System Journal, vol 28-no 4. ↗
  15. Forsberg M. and Ranta A. (2004), ‘Functional Morphology’, proceedings of the Ninth ACM SIGPLAN International Conference of Functional Programming, Snowbird, Utah. find
  16. Francopoulo G. and George M. (2008), ISO/TC 37/SC 4 N453 (N330 Rev.16). Language resource management — Lexical markup framework (LMF). find
  17. Habash N. (2004), ‘Large scale lexeme based Arabic Morphological generation’, In Proceedings of Traitement Automatique du Langage Naturel (TALN-04). Fez, Morocco. find
  18. Habash N. and Rambow O. (2005), ‘Arabic Tokenization, Part-of-Speech Tagging and Morphological Disambiguation in One Fell Swoop’, In Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics ACL 2005, pages 573 580, Ann Arbor.
  19. Habash N. and Rambow O. (2006), ‘Morphological Analysis for Arabic Dialects’, In Proceedings of COLING-ACL, Sydney, Australia. find
  20. Ibrahim K. (2002), ‘Al-Murshid fi Qawa'id Al-Nahw wa Al-Sarf’, The Guide in Syntax and Morphology Rules, Amman, Jordan: Al-Ahliyyah for Publishing and Distribution. find
  21. Khoja, S. (2001), ‘APT: Arabic Part-of-Speech Tagger’, In Proceedings of NAACL Student Research Workshop. find
  22. Koehn P. and Hoang H. (2007). ‘Factored Translation Models’. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), pages 868– 876, Prague, Czech Republic.
  23. Otakar Smrz. (2007), ‘ElixirFM. Implementation of Functional Arabic Morphology’, In ACL 2007 Proceedings of the Workshop on Computational Approaches to Semitic Languages: Common Issues and Resources, pages 1–8, Prague, Czech Republic, 2007.
  24. Tahir Y., Chenfour N., Harti M. (2004), ‘Modélisation à objets d'une base de données morphologique pour la langue arabe’, JEP-TALN 2004, Traitement Automatique de l’Arabe, Fès, Maroc. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/gridach-2011-developing-new-approach-arabic-morphological/