Elamur Research⌕
☰
home / papers / 2019 / Automatic Arabic Text Summarization Based on Fuzzy Logic

Automatic Arabic Text Summarization Based on Fuzzy Logic

Paper · 2019 · Lamees Al Qassem, Di Wang, Hassan Barada et al.

Summary

Extractive summarization for Arabic lags behind English, largely because Arabic morphology and sentence structure are hard to process. The authors build a single-document Arabic summarizer with five stages: pre-processing, noun extraction, feature extraction, fuzzy-logic scoring, and sentence selection. Sentence scoring uses only nouns, found by the team's earlier rule-based extractor that relies on Arabic grammar rather than a statistical tagger. Each sentence gets five feature values: position, TF-IDF over nouns, positive and negative cue phrases, topic-signature words, and presence of numbers. A fuzzy inference system, initialised from expert knowledge, merges these into one score. The top-scoring sentences are output in their original order, and the user picks the summary length. Evaluation on the EASC corpus with ROUGE-1 and ROUGE-2 at 10% and 30% summary sizes shows the system beating LCEAS, Al-Radaideh and Afif, Haboush et al., and Oufaida et al. The result matters because a transparent, feature-based scorer outperforms prior systems on a low-resource language.

Key points

  • The pipeline has five components: pre-processing, noun extraction, feature extraction, fuzzy logic, and sentence selection.
  • The noun extractor comes from the authors' 2018 work (Al Qassem et al., 2018). It was compared with the Stanford Arabic POS tagger and reached comparable accuracy more efficiently.
  • The five sentence features are:
    • position;
    • TF-IDF computed on nouns only;
    • cue phrases, where positive ones raise a sentence's score and negative ones lower it;
    • topic-signature words;
    • numerical data.
  • Evaluation used the EASC corpus, which has five reference summaries per document, with ROUGE-N for N = 1 and 2.
  • At the 10% size, the system's ROUGE-2 scores were recall 0.27, precision 0.40, and F-measure 0.29. For comparison, LCEAS reached an F-measure of about 0.22.

Interesting ideas

  • I1. Scoring sentence importance from nouns alone cuts noise and computation, on the premise that content-bearing ideas live in nouns.
  • I2. Cue phrases work in both directions: some signal key content, while others (such as "for example" or "in other words") mark elaboration or redundancy and should lower a sentence's score.
  • I3. Hand-crafted features are combined by a fuzzy rule system seeded with expert knowledge rather than by fixed subjective weights or an opaque classifier. This keeps each score traceable to readable rules.
  • I4. Extracted sentences are reordered back into their original document order, and the user controls output size as a percentage.
  • I5. Comparisons are made at several summary lengths (10% and 30%) against multiple references, so the reported scores are not tied to a single length setting.

For Mova

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

  • I1 (world): Entity salience in the text world could weight noun-headed nodes more heavily (UD nsubj/obj/nmod heads), giving a cheap, explainable ranking of which entities are central.
  • I2 (frames): A discourse-marker lexicon in prag could tag clauses as summary-level or elaborative, helping decide which frames to foreground when building a world.
  • I3 (answer): Answer selection that combines several soft cues could use explicit fuzzy rules, so every chosen answer comes with a human-readable justification.
  • I4 (back): The text → graph → text compressor could select salient world facts by score, then emit them in source event order with a tunable compression ratio.
  • I5 (learn): Mova's evaluation could report results across several output budgets and multiple references, exposing cases where gains appear only at one setting.

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 worksSummarization · 232 works

Cite

Cite as Lamees Al Qassem, Di Wang, Hassan Barada, Ahmad Al-Rubaie and Nawaf Almoosa. 2019. Automatic Arabic Text Summarization Based on Fuzzy Logic. 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.

@inproceedings{qassem2019automatic,
  title = {{Automatic Arabic Text Summarization Based on Fuzzy Logic}},
  author = {Lamees {Al Qassem}, Di Wang, Hassan Barada, Ahmad Al-Rubaie and Nawaf Almoosa},
  booktitle = {Proceedings of the 3rd International Conference on Natural Language and Speech Processing},
  year = {2019},
  url = {https://aclanthology.org/W19-7406/}
}

References

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

  1. F Al-Khawaldeh and V Samawi. 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
  2. Lamees Al Qassem, Di Wang, and Hassan Barada. 2018. Noun extraction tool for anlp applications. In 2018 IEEE 12th International Conference on Semantic Computing (ICSC), pages 308–309. IEEE. find
  3. Lamees Mahmoud Al Qassem, Di Wang, Zaid Al Mahmoud, Hassan Barada, Ahmad Al-Rubaie, and Nawaf I Almoosa. 2017. Automatic arabic summarization: a survey of methodologies and systems. Procedia Computer Science, 117:10–18. find
  4. Q Al-Radaideh and Mohammad Afif. 2009. Arabic text summarization using aggregate similarity. In International Arab conference on information technology (ACIT2009), Yemen. find
  5. Riyad Al-Shalabi, Ghassan Kanaan, Bashar Al-Sarayreh, Khalid Khanfar, Ali Al-Ghonmein, Hamed Talhouni, Salem Al-Azazmeh, et al. 2009. Proper noun extracting algorithm for arabic language. In International conference on IT, Thailand. find
  6. Riadh Belkebir and Ahmed Guessoum. 2015. A supervised approach to arabic text summarization using adaboost. In New Contributions in Information Systems and Technologies, pages 227–236. Springer. ↗
  7. Mohamed Mahdi Boudabous, Mohamed Hédi Maaloul, and Lamia Hadrich Belguith. 2010. Digital learning for summarizing arabic documents. In International Conference on Natural Language Processing, pages 79–84. Springer. find
  8. Dipanjan Das and André FT Martins. 2007. A survey on automatic text summarization. Literature Survey for the Language and Statistics II course at CMU, 4(192-195):57. find
  9. Fouad Soufiane Douzidia and Guy Lapalme. 2004. Lakhas, an arabic summarization system. Proceedings of DUC2004. find
  10. Harold P Edmundson. 1969. New methods in automatic extracting. Journal of the ACM (JACM), 16(2):264–285. ↗
  11. Mahmoud El-Haj, Udo Kruschwitz, and Chris Fox. 2009. Experimenting with automatic text summarisation for arabic. In Language and Technology Conference, pages 490–499. Springer. find
  12. Mahmoud El-Haj, Udo Kruschwitz, and Chris Fox. 2010. Using mechanical turk to create a corpus of arabic summaries. find
  13. Mohamed Abdel Fattah and Fuji Ren. 2009. Ga, mr, ffnn, pnn and gmm based models for automatic text summarization. Computer Speech & Language, 23(1):126–144. ↗
  14. Rafael Ferreira, Luciano de Souza Cabral, Rafael Dueire Lins, Gabriel Pereira e Silva, Fred Freitas, George DC Cavalcanti, Rinaldo Lima, Steven J Simske, and Luciano Favaro. 2013. Assessing sentence scoring techniques for extractive text summarization. Expert systems with applications, 40(14):5755–5764. ↗
  15. Ahmad Haboush, Maryam Al-Zoubi, Ahmad Momani, and Motassem Tarazi. 2012. Arabic text summarization model using clustering techniques. World of Computer Science and Information Technology Journal (WCSIT) ISSN, pages 2221–0741. find
  16. Eyke Hüllermeier. 2011. Fuzzy sets in machine learning and data mining. Applied Soft Computing, 11(2):1493–1505. find
  17. J Kupiec, J Pedersen, and F Chen. 1995. A trainable document summarizer. dans les actes de acm special interest group on information retrieval (sigir), 68– 73. find
  18. Hans Peter Luhn. 1958. The automatic creation of literature abstracts. IBM Journal of research and development, 2(2):159–165. ↗
  19. S Santhana Megala, A Kavitha, and A Marimuthu. 2014. Enriching text summarization using fuzzy logic. International Journal of Computer Science and Information Technologies, 5(1):863–867. ↗
  20. Ani Nenkova, Kathleen McKeown, et al. 2011. Automatic summarization. Foundations and Trends R in Information Retrieval, 5(2–3):103–233. find
  21. Houda Oufaida, Omar Nouali, and Philippe Blache. 2014. Minimum redundancy and maximum relevance for single and multi-document arabic text summarization. Journal of King Saud University-Computer and Information Sciences, 26(4):450– 461. ↗
  22. Mohamad Sarmini. 2015. Design and implementation of hybrid syntactic-fuzzy genetic system for extractive arabic text summarization. International Journal of Computer Science and Engineering in Arabic, 6:69–98. find
  23. Ladda Suanmali, Naomie Salim, and Mohammed Salem Binwahlan. 2009. Feature-based sentence extraction using fuzzy inference rules. In 2009 International Conference on Signal Processing Systems, pages 511–515. IEEE. find
  24. Jyoti Yadav and Yogesh Kumar Meena. 2016. Use of fuzzy logic and wordnet for improving performance of extractive automatic text summarization. In 2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), pages 2071–2077. IEEE. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/qassem-2019-automatic-arabic-text-summarization-based/