---
title: "A Dataset for Detecting Humor in Arabic Text — Alkhalifa 2022"
description: "The authors build what they present as the first Arabic humor dataset. 72% accuracy. 22 references, cited by 2 papers here."
authors: ["Hend Alkhalifa", "Fetoun Alzahrani", "Hala Qawara", "Reema Alrowais", "Sawsan Alowa", "Luluh Aldhubayi"]
year: 2022
venue: "Proceedings of the 5th International Conference on Natural Language and Speech Processing (ICNLSP 2022)"
source: "https://aclanthology.org/2022.icnlsp-1.25/"
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/alkhalifa-2022-dataset-detecting-humor-arabic-text/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/alkhalifa-2022-dataset-detecting-humor-arabic-text/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2022](https://md.research.elamur.ai/free/en/papers/2022/page/8/) / A Dataset for Detecting Humor in Arabic Text

# A Dataset for Detecting Humor in Arabic Text

**Authors:** Hend Alkhalifa, Fetoun Alzahrani, Hala Qawara, Reema Alrowais, Sawsan Alowa and Luluh Aldhubayi · **Year:** 2022 · **Venue:** [Proceedings of the 5th International Conference on Natural Language and Speech Processing (ICNLSP 2022)](https://md.research.elamur.ai/free/en/venues/icnlsp/) ([International Conference on Natural Language and Speech Processing 2022](https://md.research.elamur.ai/free/en/venues/icnlsp/2022/)) · **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/2022.icnlsp-1.25/) · [PDF](https://aclanthology.org/2022.icnlsp-1.25.pdf)

**Citations:** [2](https://md.research.elamur.ai/free/en/papers/alkhalifa-2022-dataset-detecting-humor-arabic-text/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Arabic has almost no annotated resources for recognising humor, and earlier Arabic work usually treated humor only as a side case of sarcasm detection. The authors build what they present as the first Arabic humor dataset. It covers Modern Standard Arabic and several dialects (Gulf, Egyptian, Levantine, Maghrebi). The data came from two sources: tweets scraped with Twint from accounts and hashtags known for jokes, and web text, mostly forum posts, gathered through Sketch Engine searches on keywords that typically set up a joke. After normalisation and heavy cleaning, three annotators labeled every item as humor or non-humor, following explicit guidelines for unclear or incomplete texts. To set baselines, the authors fine-tuned seven Arabic pre-trained BERT-style models as binary classifiers. The best one, CAMeLBERT-DA, reached about 72% accuracy and F1. The moderate scores show that humor detection remains hard, especially across dialects. The resource gives future Arabic humor research a shared benchmark, which the field previously lacked.

## Key points
- The corpus has 10,039 texts with 201,095 tokens and 36,814 word types. Text length ranges from 3 to 446 words. It is released publicly on GitHub (iwan-rg/Arabic-Humor).
- Class balance is close to even: 4,455 humor items (44.38%) versus 55.62% non-humor.
- Three annotators labeled each text, and inter-annotator agreement measured by Fleiss's Kappa was 0.73 ("substantial").
- The seven models were AraBERTv02, Arabertv02-twitter, QARIB, MarBERT, MARBERTv2, CAMeLBERT-DA and CAMeLBERT-MIX, evaluated on an 85:15 train/test split.
- Results:
  - CAMeLBERT-DA was best (accuracy/F1 72.11), followed by MARBERTv2 (71.71 accuracy).
  - Arabertv02-twitter and MarBERT produced identical scores.
  - QARIB was weakest (67.72 accuracy).
  - The authors attribute the gap mainly to how much dialectal data each model saw in pre-training.

## Interesting ideas
- Humor-bearing text can be harvested through "joke trigger" keywords, meaning the stock characters and set-up phrases that jokes typically open with, instead of only through labeled sources.
- The annotation guidelines settle edge cases by fixed defaults: if the annotator cannot get the joke, the label is non-humor; a tweet that lost its image but still reads as amusing is labeled humor.
- Agreement among three independent annotators is reported with Fleiss's Kappa before any model is trained, so the ceiling of the task is known in advance.
- Boilerplate from forum pages (author info, post counts, contact fields) is removed with a small dictionary of marker phrases that drops any matching row.
- Collapsing elongated letter runs ("laughter" strings) and normalising variant letter forms reduces sparsity in social-media text.

## For Mova

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

- **I1** (frames): a small lexicon of English joke openers and stock-character cues could serve as an explainable signal for the irony/humor detector in prag, flagging a text whose frames should not be read literally.
- **I2** (learn): when LLM teacher tiers or witness parsers produce silver labels for pragmatic categories, explicit default rules for uncertain or truncated inputs would make labels reproducible and keep "unsure" from leaking in as noise.
- **I3** (learn): computing a multi-rater agreement score over several witness parsers or teachers would show where silver data is trustworthy and where scores near the agreement level are already the realistic maximum.
- **I4** (tokens): a similar marker-phrase filter could strip page furniture from scraped English text before tagging, so it never reaches the parser as fake sentences.
- **I5** (tokens): mapping English elongations like "soooo" or "lolll" to a canonical lemma, while keeping a flag that elongation occurred, would help the tagger and also preserve an expressive cue for pragmatics.

*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 (2): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 2,908 works

## Cite

Cite as: Hend Alkhalifa, Fetoun Alzahrani, Hala Qawara, Reema Alrowais, Sawsan Alowa and Luluh Aldhubayi. 2022. A Dataset for Detecting Humor in Arabic Text. Proceedings of the 5th International Conference on Natural Language and Speech Processing (ICNLSP 2022).

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

```bibtex
@inproceedings{alkhalifa2022dataset,
  title = {{A Dataset for Detecting Humor in Arabic Text}},
  author = {Hend Alkhalifa, Fetoun Alzahrani, Hala Qawara, Reema Alrowais, Sawsan Alowa and Luluh Aldhubayi},
  booktitle = {Proceedings of the 5th International Conference on Natural Language and Speech Processing (ICNLSP 2022)},
  year = {2022},
  url = {https://aclanthology.org/2022.icnlsp-1.25/}
}
```

- [← previous A Context-free Arabic Emoji Sentiment Lexicon (CF-Arab-ESL)](https://md.research.elamur.ai/free/en/papers/hakami-2022-context-free-arabic-emoji-sentiment/)
- [next → aiXplain at Arabic Hate Speech 2022: An Ensemble Based Approach to Detecting Offensive Tweets](https://md.research.elamur.ai/free/en/papers/alzubi-2022-aixplain-arabic-hate-speech-2022/)
- [↑ back to papers of 2022](https://md.research.elamur.ai/free/en/papers/2022/page/8/)

## References

22 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Altin, Lutfiye Seda Mut. 2019. “LaSTUS/TALN at HAHA: Humor Analysis Based on Human Annotation.” 6. [find](https://scholar.google.com/scholar?q=Altin%2C+Lutfiye+Seda+Mut%2E+2019%2E+%E2%80%9CLaSTUS%2FTALN+at+HAHA%3A+Humor+Analysis+Based+on+Human+Annotation%2E%E2%80%9D+6%2E)
2. Amin, Miriam, and Manuel Burghardt. 2020. “A Survey on Approaches to Computational Humor Generation.” Pp. 29–41 in Proceedings of the The 4th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature. Online: International Committee on Computational Linguistics. [↗](https://doi.org/10.48448/y7rh-e163)
3. Annamoradnejad, Issa, and Gohar Zoghi. 2020. ColBERT: Using BERT Sentence Embedding for Humor Detection. [find](https://scholar.google.com/scholar?q=Annamoradnejad%2C+Issa%2C+and+Gohar+Zoghi%2E+2020%2E+ColBERT%3A+Using+BERT+Sentence+Embedding+for+Humor+Detection%2E)
4. Banikalef, Eddin Abdullah Ahmed. 2014. “Linguistic Analysis of Humor in Jordanian Arabic among Young Jordanians Facebookers.” Retrieved December 8, 2021 ([https://www.semanticscholar.org/paper/Lin](https://www.semanticscholar.org/paper/Lin) guistic-Analysis-of-Humor-in-Jordanian-Arabic-Banikalef/e6aaed4aeb4e627678b62203791d 959c792dd625). [↗](https://www.semanticscholar.org/paper/Lin)
5. Bertero, Dario, and Pascale Fung. 2016. “Multimodal Deep Neural Nets for Detecting Humor in TV Sitcoms.” Pp. 383–90 in 2016 IEEE Spoken Language Technology Workshop (SLT). [find](https://scholar.google.com/scholar?q=Bertero%2C+Dario%2C+and+Pascale+Fung%2E+2016%2E+%E2%80%9CMultimodal+Deep+Neural+Nets+for+Detecting+Humor+in+TV+Sitcoms%2E%E2%80%9D+Pp%2E+383%E2%80%9390+in+2016+IEEE+Spoken+Language+Technology+Workshop+%28SLT%29%2E)
6. Biniz, Mohamed, Samir Boukil, Fatiha Adnani, Loubna Cherrat, and Abd Moutaouakkil. 2018. “Arabic Text Classification Using Deep Learning Technics.” International Journal of Grid and Distributed Computing 11:103–14. doi: 10.14257/ijgdc.2018.11.9.09. [↗](https://doi.org/10.14257/ijgdc.2018.11.9.09)
7. Bzour, Assistant Prof Anwar Fayez Al. 2021. “Arabic Humorous Texts: An Attempt to Analyze.” Review of International Geographical Education Online 11(4):1480-1492. [find](https://scholar.google.com/scholar?q=Bzour%2C+Assistant+Prof+Anwar+Fayez+Al%2E+2021%2E+%E2%80%9CArabic+Humorous+Texts%3A+An+Attempt+to+Analyze%2E%E2%80%9D+Review+of+International+Geographical+Education+Online+11%284%29%3A1480%2D1492%2E)
8. Castro, Santiago, Matías Cubero, Diego Garat, and Guillermo Moncecchi. 2016. “Is This a Joke? Detecting Humor in Spanish Tweets.” ArXiv:1703.09527 [Cs] 10022:139–50. doi: 10.1007/978-3-319-47955-2_12. [↗](https://doi.org/10.1007/978-3-319-47955-2_12)
9. Fan, Xiaochao, Hongfei Lin, Liang Yang, Yufeng Diao, Chen Shen, Yonghe Chu, and Yanbo Zou. 2020. “Humor Detection via an Internal and External Neural Network.” Neurocomputing 394. doi: 10.1016/j.neucom.2020.02.030. [↗](https://doi.org/10.1016/j.neucom.2020.02.030)
10. Fleiss, Joseph L. 1971. “Measuring Nominal Scale Agreement among Many Raters.” Psychological Bulletin 76(5):378–82. doi: 10.1037/h0031619. [↗](https://doi.org/10.1037/h0031619)
11. Hasan, Md Kamrul, Wasifur Rahman, Amir Zadeh, Jianyuan Zhong, Md Iftekhar Tanveer, Louis-Philippe Morency, Mohammed, and Hoque. 2019. “UR-FUNNY: A Multimodal Language Dataset for Understanding Humor.” Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) 2046–56. doi: 10.18653/v1/D19-1211. [↗](https://doi.org/10.18653/v1/d19-1211)
12. Hossain, Nabil, John Krumm, and Michael Gamon. 2019. “‘President Vows to Cut \<Taxes> Hair’: Dataset and Analysis of Creative Text Editing for Humorous Headlines.” ArXiv:1906.00274 [Cs]. [↗](https://arxiv.org/abs/1906.00274)
13. Kao, Justine T., R. Levy, and Noah D. Goodman. 2016. “A Computational Model of Linguistic Humor in Puns.” Cogn. Sci. doi: 10.1111/cogs.12269. [↗](https://doi.org/10.1111/cogs.12269)
14. Khandelwal, Ankush, Sahil Swami, Syed S. Akhtar, and Manish Shrivastava. 2018. “Humor Detection in English-Hindi Code-Mixed Social Media Content : Corpus and Baseline System.” ArXiv:1806.05513 [Cs]. [↗](https://arxiv.org/abs/1806.05513)
15. Landis, J. Richard, and Gary G. Koch. 1977. “The Measurement of Observer Agreement for Categorical Data.” Biometrics 33(1):159. doi: 10.2307/2529310. [↗](https://doi.org/10.2307/2529310)
16. Li, Da, Rafal Rzepka, Michal Ptaszynski, and Kenji Araki. 2020. “HEMOS: A Novel Deep Learning-Based Fine-Grained Humor Detecting Method for Sentiment Analysis of Social Media.” Information Processing & Management 57:102290. doi: 10.1016/j.ipm.2020.102290. [↗](https://doi.org/10.1016/j.ipm.2020.102290)
17. Meaney, J. A., Steven Wilson, Luis Chiruzzo, Adam Lopez, and Walid Magdy. 2021. “[SemEval 2021 Task 7: HaHackathon, Detecting and Rating Humor and Offense](https://md.research.elamur.ai/free/en/papers/meaney-2021-semeval-2021-task-7-hahackathon/).” Pp. 105–19 in Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021). Online: Association for Computational Linguistics.
18. Meyer, John C. 2000. “Humor as a Double-Edged Sword: Four Functions of Humor in Communication.” Communication Theory 10(3):310–31. doi: 10.1111/j.1468-2885.2000.tb00194.x. [↗](https://doi.org/10.1111/j.1468-2885.2000.tb00194.x)
19. Nayef, Heba, and Mohamed El-Nashar. 2014. “‘Dissecting the Poisoned Honey’ Sexist Humor in Egypt: A Linguistic Analysis of Sexism in Colloquial Cairene Arabic Jokes.” Anàlisi 131. doi: 10.7238/a.v0i50.2324. [↗](https://doi.org/10.7238/a.v0i50.2324)
20. Omar, Hidaya Moulay. 2017. “A Linguistic Analysis of Humour in Algerian Arabic : The Case of ‘Sultan Ashur X’ Sitcom.” Thesis, Yarmouk University. [find](https://scholar.google.com/scholar?q=Omar%2C+Hidaya+Moulay%2E+2017%2E+%E2%80%9CA+Linguistic+Analysis+of+Humour+in+Algerian+Arabic+%3A+The+Case+of+%E2%80%98Sultan+Ashur+X%E2%80%99+Sitcom%2E%E2%80%9D+Thesis%2C+Yarmouk+University%2E)
21. Zhang, Dongyu, Heting Zhang, Xikai Liu, Hongfei Lin, and Feng Xia. 2019. “Telling the Whole Story: A Manually Annotated Chinese Dataset for the Analysis of Humor in Jokes.” Pp. 6402–7 in Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). Hong Kong, China: Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Zhang%2C+Dongyu%2C+Heting+Zhang%2C+Xikai+Liu%2C+Hongfei+Lin%2C+and+Feng+Xia%2E+2019%2E+%E2%80%9CTelling+the+Whole+Story%3A+A+Manually+Annotated+Chinese+Dataset+for+the+Analysis+of+Humor+in+Jokes%2E%E2%80%9D+Pp%2E+6402%E2%80%937+in+Proceedings+o)
22. Zhang, Renxian, and Naishi Liu. 2014. “Recognizing Humor on Twitter.” Pp. 889–98 in Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management, CIKM ’14. New York, NY, USA: Association for Computing Machinery. [find](https://scholar.google.com/scholar?q=Zhang%2C+Renxian%2C+and+Naishi+Liu%2E+2014%2E+%E2%80%9CRecognizing+Humor+on+Twitter%2E%E2%80%9D+Pp%2E+889%E2%80%9398+in+Proceedings+of+the+23rd+ACM+International+Conference+on+Conference+on+Information+and+Knowledge+Management%2C+CIKM+%E2%80%9914)

*Updated 1 November 2026*