---
title: "SumSAT — paper by Cheragui and Lakhdar (2019)"
description: "SumSAT is an extractive summarizer for Arabic, a language where tools lag behind English because its morphology and… 19 references."
authors: ["Mohamed Amine Cheragui", "Said Moulay Lakhdar"]
year: 2019
venue: "Proceedings of the 3rd International Conference on Natural Language and Speech Processing"
source: "https://aclanthology.org/W19-7417/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/cheragui-2019-sumsat-hybrid-arabic-text-summarization/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/cheragui-2019-sumsat-hybrid-arabic-text-summarization/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2019](https://md.research.elamur.ai/free/en/papers/2019/page/9/) / SumSAT: Hybrid Arabic Text Summarization based on symbolic and numerical Approaches

# SumSAT: Hybrid Arabic Text Summarization based on symbolic and numerical Approaches

**Authors:** [Mohamed Amine Cheragui](https://md.research.elamur.ai/free/en/authors/mohamed-amine-cheragui/) and Said Moulay Lakhdar · **Year:** 2019 · **Venue:** [Proceedings of the 3rd International Conference on Natural Language and Speech Processing](https://md.research.elamur.ai/free/en/venues/icnlsp/) ([International Conference on Natural Language and Speech Processing 2019](https://md.research.elamur.ai/free/en/venues/icnlsp/2019/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/W19-7417/) · [PDF](https://aclanthology.org/W19-7417.pdf) · [OpenAlex](https://openalex.org/W2990597844)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
SumSAT is an extractive summarizer for Arabic, a language where tools lag behind English because its morphology and syntax are complex. The system has three stages. Pre-processing splits the text into sentences with rule-based contextual exploration over punctuation and reduces words to roots with a Full-Text Search facility, which also drops stopwords. Processing scores sentences in three ways:
- The main, symbolic method is contextual exploration: hand-written rules look for an indicator and then confirm it with nearby clues before tagging a sentence with one of 13 discursive categories.
- A secondary method marks sentences that contain domain-specific indicators.
- A second secondary method is a graph model whose node weights come from roots shared between sentences, traversed greedily.

A final filtering step applies the user's chosen viewpoint and compression ratio. The evaluation covers only the category annotation. On 25 texts with reference summaries written by two experts, precision and recall exceed 66% for most categories, though a few categories fall well below that. The work matters because it lets users request a summary from a chosen viewpoint and keeps a statistical summary as a fallback.

## Key points
- The system defines 13 discursive categories, for example Objective, Hypothesis, Explanation, Consequence, Definition, Confirmation, Reminder, Prediction and Enunciation. Each category has its own indicators and complementary clues.
- Rules are stored in a tagged format with two parts:
  - **Conditions:** the indicator, the clues, the search space (sentence or paragraph) and the context (before or after the indicator).
  - **Actions:** for example, labelling a sentence "Conclusion".
- The evaluation set has 25 texts with human reference summaries. The reported scores include:
  - Explanation: P 88.57 / R 95.38
  - Confirmation: P 97.5 / R 82.98
  - Enunciation: P 94.94 / R 91.85
  - Definition: recall only 32.67
- Hypothesis, Recapitulation, Reminder and Prediction reach only 40–50% precision. Prediction, Definition and Reminder have recall between 30% and 50%.
- The authors list three error sources:
  - Authors' irregular punctuation causes segmentation errors.
  - Collecting markers by hand is costly.
  - The greedy graph walk can skip heavy sentences because it only moves between adjacent nodes.

## Interesting ideas
- An indicator alone never fires a label: a rule must also find a confirming clue in a declared window and direction. This cuts false positives from ambiguous trigger words.
- Sentence splitting is treated as a rule-driven decision at each punctuation mark, with explicit exceptions for abbreviations, acronyms and decimals, rather than a fixed split on every period.
- The design is layered: a precise symbolic method runs first, and a cheap statistical method is used only when the symbolic layer finds nothing. This lets the system always produce output while staying explainable when the rules do apply.
- Scores are reported per category against expert references, and the authors separate low precision (needs better marker patterns) from low recall (needs more markers). This turns the error analysis into concrete next steps.
- Sentence importance is measured as the number of shared roots with all other sentences, a simple and transparent centrality measure.

## For Mova

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

- **I1** (frames): Speech-act and implicature detection in prag could require a trigger plus a positioned corroborating cue, giving each label an explainable two-part justification.
- **I2** (tokens): The en tokenizer's sentence boundaries could be expressed as inspectable indicator+context rules, so each split or non-split can be traced to a named rule.
- **I3** (back): The text→graph→text compressor could select content by frame or discourse role first and fall back to overlap-based salience when no role markers are present.
- **I4** (learn): Evaluations of frame or discourse labels could be broken down per label, with precision gaps routed to rule repair and recall gaps routed to silver-data or lexicon expansion.
- **I5** (world): Overlap counts of entities or lemmas across sentences could rank which events and entities belong at the core of a text's world, with the counts serving as the explanation.

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

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [📝 Summarization](https://md.research.elamur.ai/free/en/topics/summarization/) · 243 works

## Cite

Cite as: Mohamed Amine Cheragui and Said Moulay Lakhdar. 2019. SumSAT: Hybrid Arabic Text Summarization based on symbolic and numerical Approaches. Proceedings of the 3rd International Conference on Natural Language and Speech Processing.

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

```bibtex
@inproceedings{cheragui2019sumsat,
  title = {{SumSAT: Hybrid Arabic Text Summarization based on symbolic and numerical Approaches}},
  author = {Mohamed Amine Cheragui and Said Moulay Lakhdar},
  booktitle = {Proceedings of the 3rd International Conference on Natural Language and Speech Processing},
  year = {2019},
  url = {https://aclanthology.org/W19-7417/}
}
```

- [← previous ST NSURL 2019 Shared Task: Semantic Question Similarity in Arabic](https://md.research.elamur.ai/free/en/papers/lichouri-2019-st-nsurl-2019-shared-task/)
- [next → Teaching English Pragmatics by Taking a Sociocultural Stance](https://md.research.elamur.ai/free/en/papers/zaferanieh-2019-teaching-english-pragmatics-taking-sociocultural/)
- [↑ back to papers of 2019](https://md.research.elamur.ai/free/en/papers/2019/page/9/)

## References

19 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. study. In Proceedings of the 11th International Conference on Computer Engineering & Systems (ICCES). Egypt Allahyari Mehdi , Pouriyeh Seyedamin , Assefi Mehdi, Safaei Saeid , Trippe D. Elizabeth, Gutierrez B.Juan and Kochut Krys. 2017. Text Summarization Techniques: A Brief Survey. In Proceedings of arXiv. rXiv:1707.02268. [find](https://scholar.google.com/scholar?q=study%2E+In+Proceedings+of+the+11th+International+Conference+on+Computer+Engineering+%26+Systems+%28ICCES%29%2E+Egypt+Allahyari+Mehdi+%2C+Pouriyeh+Seyedamin+%2C+Assefi+Mehdi%2C+Safaei+Saeid+%2C+Trippe+D%2E+Elizabeth%2C+Gut)
2. Munot Nikita and Govilkar S. Sharvari. 2014 . Comparative Study of Text Summarization Methods. International Journal of Computer Applications. Volume 102–No.12. [find](https://scholar.google.com/scholar?q=Munot+Nikita+and+Govilkar+S%2E+Sharvari%2E+2014+%2E+Comparative+Study+of+Text+Summarization+Methods%2E+International+Journal+of+Computer+Applications%2E+Volume+102%E2%80%93No%2E12%2E)
3. Alkhawaldeh Fatima Taha and Samawi W. Venus. 2015. Lexical cohesion and entailment based segmentation for arabic text summarization (lceas). The World of Computer Science and Information Technology Journal. (WSCIT)5 (3): 51-60. [find](https://scholar.google.com/scholar?q=Alkhawaldeh+Fatima+Taha+and+Samawi+W%2E+Venus%2E+2015%2E+Lexical+cohesion+and+entailment+based+segmentation+for+arabic+text+summarization+%28lceas%29%2E+The+World+of+Computer+Science+and+Information+Technology+Jo)
4. Oufaida Houda., Noualib Omar and Blache Philippe. 2014. Minimum redundancy and maximum relevance for single and multi-document Arabic text summarization. Journal of King Saud University -Computer and Information Sciences, Volume 26, Issue 4. [find](https://scholar.google.com/scholar?q=Oufaida+Houda%2E%2C+Noualib+Omar+and+Blache+Philippe%2E+2014%2E+Minimum+redundancy+and+maximum+relevance+for+single+and+multi%2Ddocument+Arabic+text+summarization%2E+Journal+of+King+Saud+University+%2DComputer+and+)
5. Alotaiby Fahad, Foda Salah and Alkharashi Ibrahim. 2012. New approaches to automatic headline generation for Arabic documents. Journal of Engineering and Computer Innovations. Vol. 3(1), pp. 11-25. [find](https://scholar.google.com/scholar?q=Alotaiby+Fahad%2C+Foda+Salah+and+Alkharashi+Ibrahim%2E+2012%2E+New+approaches+to+automatic+headline+generation+for+Arabic+documents%2E+Journal+of+Engineering+and+Computer+Innovations%2E+Vol%2E+3%281%29%2C+pp%2E+11%2D25%2E)
6. Pal Alok Ranjan and Saha Diganta. 2014. An Approach to Automatic Text Summarization using WordNet. In Proceedings of the 4 th Advance Computing Conference (IACC), Page(s): 1169 – 1173. [find](https://scholar.google.com/scholar?q=Pal+Alok+Ranjan+and+Saha+Diganta%2E+2014%2E+An+Approach+to+Automatic+Text+Summarization+using+WordNet%2E+In+Proceedings+of+the+4+th+Advance+Computing+Conference+%28IACC%29%2C+Page%28s%29%3A+1169+%E2%80%93+1173%2E)
7. India Belguith Lamia Hadrich. 2014. Automatic summarization. Natural Language Processing of Semitic Languages. Springer Berlin Heidelberg.371-408. [find](https://scholar.google.com/scholar?q=India+Belguith+Lamia+Hadrich%2E+2014%2E+Automatic+summarization%2E+Natural+Language+Processing+of+Semitic+Languages%2E+Springer+Berlin+Heidelberg%2E371%2D408%2E)
8. Pai Anusha. 2014. Text Summarizer Using Abstractive and Extractive Method. International Journal of Engineering Research & Technology, Vol. 3 Issue 5. [find](https://scholar.google.com/scholar?q=Pai+Anusha%2E+2014%2E+Text+Summarizer+Using+Abstractive+and+Extractive+Method%2E+International+Journal+of+Engineering+Research+%26+Technology%2C+Vol%2E+3+Issue+5%2E)
9. Belkebir Riadh. and Guessoum Ahmed. 2015. A supervised approach to arabic text summarization using adaboost. New Contributions in Information Systems and Technologies. Springer International Publishing. 227-236. [find](https://scholar.google.com/scholar?q=Belkebir+Riadh%2E+and+Guessoum+Ahmed%2E+2015%2E+A+supervised+approach+to+arabic+text+summarization+using+adaboost%2E+New+Contributions+in+Information+Systems+and+Technologies%2E+Springer+International+Publishin)
10. Rouibia Rima., Belhadj Imane. & Cheragui Mohamed Amine. 2017. JIDR: Towards building hybrid Arabic stemmer. In Proceeding of the 1 st IEEE International Conference on Mathematics and Information Technology (ICMIT). Adrar. Algeria. [find](https://scholar.google.com/scholar?q=Rouibia+Rima%2E%2C+Belhadj+Imane%2E+%26+Cheragui+Mohamed+Amine%2E+2017%2E+JIDR%3A+Towards+building+hybrid+Arabic+stemmer%2E+In+Proceeding+of+the+1+st+IEEE+International+Conference+on+Mathematics+and+Information+Techn)
11. D'Avanzo Ernesto, Magnini Bernardo and Vallin Alessandro. 2004. Keyphrase Extraction for Summarization Purposes: The LAKE System at DUC-2004. In Proceedings of the 2004 Document Understanding Conference (DUC2004), Boston, MA. [find](https://scholar.google.com/scholar?q=D%27Avanzo+Ernesto%2C+Magnini+Bernardo+and+Vallin+Alessandro%2E+2004%2E+Keyphrase+Extraction+for+Summarization+Purposes%3A+The+LAKE+System+at+DUC%2D2004%2E+In+Proceedings+of+the+2004+Document+Understanding+Conferen)
12. Schlesinger D. Judith , O’Leary P. Dianne and Conroy M. John 2008. Arabic/English multi-document summarization with CLASSY—the past and the future. In Proceedings of the International Conference on Intelligent Text Processing and Computational Linguistics. Springer Berlin Heidelberg. 568-581, Douzidia Fouad Sofiane and Lapalme Guy. 2004. Lakhas, an Arabic Summarization System. In Proceedings of the 2004 Document Understanding Conference (DUC2004), Boston, MA. [find](https://scholar.google.com/scholar?q=Schlesinger+D%2E+Judith+%2C+O%E2%80%99Leary+P%2E+Dianne+and+Conroy+M%2E+John+2008%2E+Arabic%2FEnglish+multi%2Ddocument+summarization+with+CLASSY%E2%80%94the+past+and+the+future%2E+In+Proceedings+of+the+International+Conference+on+In)
13. Sobh Ibrahim , Darwish Nevin Mahmoud , Fayek Magda B. 2006. An Optimized Dual Classification System for Arabic Extractive Generic Text Summarization. Available at: [http://www.rdi-eg.com/rdi/technologies/papers.htm](http://www.rdi-eg.com/rdi/technologies/papers.htm). [↗](http://www.rdi-eg.com/rdi/technologies/papers.htm)
14. Lagrini Samira, Redjimi Mohamme and Azizi Nabiha. 2017. Extractive Arabic Text Summarization Approaches. In Proceeding of the 6th International Conference, ICALP 2017, Fez, Morocco. [find](https://scholar.google.com/scholar?q=Lagrini+Samira%2C+Redjimi+Mohamme+and+Azizi+Nabiha%2E+2017%2E+Extractive+Arabic+Text+Summarization+Approaches%2E+In+Proceeding+of+the+6th+International+Conference%2C+ICALP+2017%2C+Fez%2C+Morocco%2E)
15. Zhu Junyan, Wang Can, He Xiaofei, Bu Jiajun, Chen Cun, Shang Shujie, Qu Mingcheng and Lu Gang. 2009. Tagoriented document summarization. In Proceedings of the 18 th International Conference on World Wide Web, WWW ’09, ACM, New York, NY, USA. [find](https://scholar.google.com/scholar?q=Zhu+Junyan%2C+Wang+Can%2C+He+Xiaofei%2C+Bu+Jiajun%2C+Chen+Cun%2C+Shang+Shujie%2C+Qu+Mingcheng+and+Lu+Gang%2E+2009%2E+Tagoriented+document+summarization%2E+In+Proceedings+of+the+18+th+International+Conference+on+World+W)
16. Luhn Hans Peter. 1958. The automatic creation of literature abstracts. IBM Journal of Research and Development, Volume 2 Issue 2. [find](https://scholar.google.com/scholar?q=Luhn+Hans+Peter%2E+1958%2E+The+automatic+creation+of+literature+abstracts%2E+IBM+Journal+of+Research+and+Development%2C+Volume+2+Issue+2%2E)
17. Maâloul Mohame Hedi. 2012. Approche hybride pour le résumé automatique de textes. Application à la langue arabe. thesis doctorat, University Aix-Marseille. [find](https://scholar.google.com/scholar?q=Ma%C3%A2loul+Mohame+Hedi%2E+2012%2E+Approche+hybride+pour+le+r%C3%A9sum%C3%A9+automatique+de+textes%2E+Application+%C3%A0+la+langue+arabe%2E+thesis+doctorat%2C+University+Aix%2DMarseille%2E)
18. Mahmoud El hadj, Kruschwitz Udo and Fox Chris. 2009. Experimenting with automatic text summarisation for Arabic. Language and Technology Conference. Springer Berlin Heidelberg. 490-499. [find](https://scholar.google.com/scholar?q=Mahmoud+El+hadj%2C+Kruschwitz+Udo+and+Fox+Chris%2E+2009%2E+Experimenting+with+automatic+text+summarisation+for+Arabic%2E+Language+and+Technology+Conference%2E+Springer+Berlin+Heidelberg%2E+490%2D499%2E)
19. Mohamed Ashraf Ali. 2016. Automatic summarization of the Arabic documents using NMF: A preliminary [find](https://scholar.google.com/scholar?q=Mohamed+Ashraf+Ali%2E+2016%2E+Automatic+summarization+of+the+Arabic+documents+using+NMF%3A+A+preliminary)

*Updated 2 November 2026*