---
title: "Identification of Naturally Occurring Numerical Expressions in"
description: "First, they build a gold standard semi-automatically by projecting English number spans onto Arabic through… 94.8% F-score. 14 references."
authors: ["Nizar Habash", "Ryan Roth"]
year: 2008
venue: "Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)"
source: "https://aclanthology.org/L08-1541/"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/habash-2008-identification-naturally-occurring-numerical-expressions/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/habash-2008-identification-naturally-occurring-numerical-expressions/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2008](https://md.research.elamur.ai/free/en/papers/2008/page/2/) / Identification of Naturally Occurring Numerical Expressions in Arabic

# Identification of Naturally Occurring Numerical Expressions in Arabic

**Authors:** [Nizar Habash](https://md.research.elamur.ai/free/en/authors/nizar-habash/) and [Ryan Roth](https://md.research.elamur.ai/free/en/authors/ryan-roth/) · **Year:** 2008 · **Venue:** [Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)](https://md.research.elamur.ai/free/en/venues/lrec/) ([LREC 2008](https://md.research.elamur.ai/free/en/venues/lrec/2008/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/L08-1541/) · [PDF](https://aclanthology.org/L08-1541.pdf) · [OpenAlex](https://openalex.org/W30061678)

**Citations:** 6 citations (OpenAlex, 2026-10-14)

## Summary
The paper tackles Number Identification (NUMID) in running text. The task has two parts: finding the span of a numerical expression and converting it to a normalized digit-based value. Earlier work mostly converted isolated number words to digits, did so out of context, and broke on small format variations. Arabic adds further difficulty: numbers carry clitics and inflect for gender, definiteness and case, text usually lacks diacritics, and many words are ambiguous between numeric and non-numeric readings. The authors make two contributions. First, they build a gold standard semi-automatically by projecting English number spans onto Arabic through the hand-aligned IBM Arabic-English corpus (IBMAC), then review the result manually. Second, they present a rule-based Arabic system with two stages, SpanID and NumNorm, both built on a dedicated number lexicon. It covers digits, number words, mixed forms, ordinals, plurals and fractions. On a blind test the system reaches 94.8% F-score for spans and 92.1% for core values. The projection approach itself does not depend on the language pair, which makes evaluation data cheap to produce.

## Key points
- Gold data comes from the afa.align and treebank.align subsets of IBMAC (8818 sentences), which are split into devtest and test. The corpus's reported inter-annotator alignment agreement is about 93%.
- Manual review changed fewer than 10% of the automatic projection decisions. A corrected 400-sentence sample (370 NUMs) showed the gold to have 100% precision and 98.4% recall (99.2% F-1).
- Evaluation uses three levels, each an upper bound on the next: SpanID, Core-Match (core value) and Full-Match (value plus extras such as ordinality). On devtest, Core-Match reached 95.1 precision and 96.7 recall.
- The error analysis attributes errors either to the gold standard (e.g., mistranslated values such as 1700→17,000) or to the system. Among system errors, spurious NUMs (e.g., the names "Powell" and "Thani") outnumber missed ones by more than 4 to 1.
- The prior work critiqued includes Bringert's GF-based Numeral Translator, which fails on simple variants like "twenty-five thousand and six". MUC NUMEX is also discussed: it covers only money and percentages and does no normalization.

## Interesting ideas
- Build evaluation labels for a hard language by tagging the easy side of a word-aligned parallel corpus and carrying the spans across, then drop pairs whose projected span picks up non-target words.
- Validate a semi-automatic gold set by hand-correcting a small sample and reporting precision and recall of the gold itself before trusting it.
- Split every error into "gold was wrong" versus "system was wrong" and sub-type each (wrong value, wrong span, missed, added).
- Separate span detection from value normalization, and score them at nested levels where each level bounds the next.
- Accept mixed forms (digits combined with scale words) and common spelling variants by design, rather than keeping a small strict numeral grammar.

## For Mova

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

- **I1** (learn): We could create silver number or measure annotations for English from aligned resources in other languages, and use the filtering rule as a cheap noise control.
- **I2** (learn): Silver data agreed by our witness parsers could get the same audited-sample check, which gives an explicit error budget for the training data.
- **I3** (answer): Failed school-test or word-problem items could be triaged this way, so key errors are kept apart from reasoning errors.
- **I4** (world): Number and measure handling could first mark the span in the UD tree, then compile a value with an explainable trace, and each step could be scored on its own.
- **I5** (tokens): The en tokenizer and lemmatizer could normalize forms like "2.8 million" or "twenty five thousand and six" into one value-bearing token group for later stages.

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

## Institutions

- [Columbia University](https://md.research.elamur.ai/free/en/universities/columbia-university/) — University, New York, United States · 149 works · 48,852 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 1,282 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 1,484 works

## Cite

Cite as: Nizar Habash and Ryan Roth. 2008. Identification of Naturally Occurring Numerical Expressions in Arabic. Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{habash2008identification,
  title = {{Identification of Naturally Occurring Numerical Expressions in Arabic}},
  author = {Nizar Habash and Ryan Roth},
  booktitle = {Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)},
  year = {2008},
  doi = {10.63317/3nhyrb4uvaiz},
  url = {https://doi.org/10.63317/3nhyrb4uvaiz}
}
```

- [← previous Construct State Modification in the Arabic Treebank](https://md.research.elamur.ai/free/en/papers/gabbard-2008-construct-state-modification-arabic-treebank/)
- [next → Unsupervised Learning-based Anomalous Arabic Text Detection](https://md.research.elamur.ai/free/en/papers/abouzakhar-2008-unsupervised-learning-based-anomalous-arabic/)
- [↑ back to papers of 2008](https://md.research.elamur.ai/free/en/papers/2008/page/2/)

## References

14 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Fawaz S. Al-Anzi. 2001. Sentential Count Rules for Arabic Language. Computers and the Humanities, 35(2):153–166. [↗](https://doi.org/10.1023/a:1002433904577)
2. Björn Bringert. 2004. Numeral Translator. [http://www.cs.chalmers.se/˜bringert/gf/translate/](http://www.cs.chalmers.se/˜bringert/gf/translate/). [↗](http://www.cs.chalmers.se/˜bringert/gf/translate/)
3. Tim Buckwalter. 2002. Buckwalter Arabic morphological analyzer version 1.0. Linguistic Data Consortium, University of Pennsylvania. LDC Catalog No.: LDC2002L49. [↗](https://doi.org/10.35111/7vzm-mb15)
4. Ali Dada. 2007. Implementation of Arabic numerals and their syntax in GF. In Proceedings of the 2007 Workshop on Computational Approaches to Semitic Languages, pages 9–16, Prague, Czech Republic. [find](https://scholar.google.com/scholar?q=Ali+Dada%2E+2007%2E+Implementation+of+Arabic+numerals+and+their+syntax+in+GF%2E+In+Proceedings+of+the+2007+Workshop+on+Computational+Approaches+to+Semitic+Languages%2C+pages+9%E2%80%9316%2C+Prague%2C+Czech+Republic%2E)
5. Mona Diab, Kadri Hacioglu, and Dan Jurafsky. 2004. [Automatic Tagging of Arabic Text: From Raw Text to Base Phrase Chunks](https://md.research.elamur.ai/free/en/papers/diab-2004-automatic-tagging-arabic-text-raw/). In Proceedings of the North American Chapter of the Association for Computational Linguistics (NAACL), Boston, MA.
6. Benjamin Farber, Dayne Freitag, Nizar Habash and Owen Rambow. 2008. Improving NER in Arabic Using a Mor-phhological Tagger. In Proceedings of Language Resources and Evaluation Conference (LREC), Marrakech, Morocco. [find](https://scholar.google.com/scholar?q=Benjamin+Farber%2C+Dayne+Freitag%2C+Nizar+Habash+and+Owen+Rambow%2E+2008%2E+Improving+NER+in+Arabic+Using+a+Mor%2Dphhological+Tagger%2E+In+Proceedings+of+Language+Resources+and+Evaluation+Conference+%28LREC%29%2C+Marra)
7. Ralph Grishman and Beth Seundheim. 1996. Design of the MUC-6 evaluation. In TIPSTER Text Program Phase II Workshop, Vienna, VA. [find](https://scholar.google.com/scholar?q=Ralph+Grishman+and+Beth+Seundheim%2E+1996%2E+Design+of+the+MUC%2D6+evaluation%2E+In+TIPSTER+Text+Program+Phase+II+Workshop%2C+Vienna%2C+VA%2E)
8. Nizar Habash, Abdelhadi Soudi and Tim Buckwalter. 2007. On Arabic Transliteration. In A. van den Bosch and A. Soudi, editors. Arabic Computational Morphology: Knowledge-based and Empirical Methods, Springer, 2007. [↗](https://doi.org/10.1007/978-1-4020-6046-5_2)
9. Nizar Habash and Owen Rambow. 2005. Tokenization, Morphological Analysis, and Part-of-Speech Tagging for Arabic in One Fell Swoop. In Proceedings of Association for Computational Linguistics (ACL), Ann Arbor, Michigan. [find](https://scholar.google.com/scholar?q=Nizar+Habash+and+Owen+Rambow%2E+2005%2E+Tokenization%2C+Morphological+Analysis%2C+and+Part%2Dof%2DSpeech+Tagging+for+Arabic+in+One+Fell+Swoop%2E+In+Proceedings+of+Association+for+Computational+Linguistics+%28ACL%29%2C+An)
10. Abraham Ittycheriah and Salim Roukos. 2005. A maximum entropy word aligner for Arabic-English machine translation. In Proceedings of Empirical Methods in Natural Language Processing Conference, pages 89–96, Vancouver, Canada. [find](https://scholar.google.com/scholar?q=Abraham+Ittycheriah+and+Salim+Roukos%2E+2005%2E+A+maximum+entropy+word+aligner+for+Arabic%2DEnglish+machine+translation%2E+In+Proceedings+of+Empirical+Methods+in+Natural+Language+Processing+Conference%2C+pages+)
11. Lauri Karttunen. 2006. Numbers and Finnish Numerals. SKY Journal of Linguistics, 19:407–421. [find](https://scholar.google.com/scholar?q=Lauri+Karttunen%2E+2006%2E+Numbers+and+Finnish+Numerals%2E+SKY+Journal+of+Linguistics%2C+19%3A407%E2%80%93421%2E)
12. Bengt Sigurd. 1973. From numbers to numerals and vice versa. In Proceedings of the International Conference on Computational Linguistics, Pisa, Italy. [find](https://scholar.google.com/scholar?q=Bengt+Sigurd%2E+1973%2E+From+numbers+to+numerals+and+vice+versa%2E+In+Proceedings+of+the+International+Conference+on+Computational+Linguistics%2C+Pisa%2C+Italy%2E)
13. Richard Sproat. 2000. Lextools: a toolkit for finite-state linguistic analysis. [http://www.research.att.com/˜alb/lextools/synth.pdf](http://www.research.att.com/˜alb/lextools/synth.pdf). [↗](http://www.research.att.com/˜alb/lextools/synth.pdf)
14. David Yarowsky, Grace Ngai, and Richard Wicentowski. 2001. [Inducing multilingual text analysis tools via robust projection across aligned corpora](https://md.research.elamur.ai/free/en/papers/yarowsky-2001-inducing-multilingual-text-analysis-tools/). In Proceedings of Human Language Technology Conference, pages 161–168.

*Updated 24 October 2026*