---
title: "Arabic POS Tagging: Don't Abandon Feature Engineering Just Yet"
description: "The authors build two taggers that use the simplified 18-tag Farasa tagset and train them on ATB… 12 references, cited by 14 papers here."
authors: ["Kareem Darwish", "Hamdy Mubarak", "Ahmed Abdelali", "Mohamed Eldesouki"]
year: 2017
venue: "Proceedings of the Third Arabic Natural Language Processing Workshop"
source: "https://aclanthology.org/W17-1316/"
doi: "10.18653/v1/w17-1316"
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/darwish-2017-arabic-pos-tagging-don-t/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/darwish-2017-arabic-pos-tagging-don-t/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2017](https://md.research.elamur.ai/free/en/papers/2017/page/4/) / Arabic POS Tagging: Don't Abandon Feature Engineering Just Yet

# Arabic POS Tagging: Don't Abandon Feature Engineering Just Yet

**Authors:** [Kareem Darwish](https://md.research.elamur.ai/free/en/authors/kareem-darwish/), [Hamdy Mubarak](https://md.research.elamur.ai/free/en/authors/hamdy-mubarak/), [Ahmed Abdelali](https://md.research.elamur.ai/free/en/authors/ahmed-abdelali/) and [Mohamed Eldesouki](https://md.research.elamur.ai/free/en/authors/mohamed-eldesouki/) · **Year:** 2017 · **Venue:** [Proceedings of the Third Arabic Natural Language Processing Workshop](https://md.research.elamur.ai/free/en/venues/wanlp/) ([WANLP 2017](https://md.research.elamur.ai/free/en/venues/wanlp/2017/)) · **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/W17-1316/) · [PDF](https://aclanthology.org/W17-1316.pdf) · [DOI 10.18653/v1/w17-1316](https://doi.org/10.18653/v1/w17-1316) · [OpenAlex](https://openalex.org/W2740918723)

**Citations:** 33 citations (OpenAlex, 2026-10-14) · [14](https://md.research.elamur.ai/free/en/papers/darwish-2017-arabic-pos-tagging-don-t/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The paper asks whether hand-crafted linguistic features are still useful for Arabic part-of-speech tagging now that bi-LSTM taggers with embeddings can learn representations directly from data. The authors build two taggers that use the simplified 18-tag Farasa tagset and train them on ATB parts 1–3. The first is an SVM Rank tagger with a large set of probabilistic features: clitic and word context, affixes, stem templates, meta types and a verb-likelihood lexicon. It can tag at the clitic level or assign a compound tag to the whole word. The second is a clitic-level bi-LSTM, tested with and without these features and with embeddings trained on Aljazeera.net text. Both are compared with MADAMIRA on the multi-genre WikiNews test set, using either gold or Farasa segmentation. Explicit features and embeddings each help the bi-LSTM, but features help more and hold up better when segmentation is noisy. Combining the two makes results worse. The best bi-LSTM only slightly beats SVM Rank. The work suggests that linguistic knowledge still adds value over purely learned representations, and both taggers are released as open source.

## Key points
- Training data is ATB parts 1 (v4.1), 2 (v3.1) and 3 (v2). The test set is WikiNews: 70 articles from 2013–2014, about 18,300 words, manually segmented and tagged, covering politics, economics, health, science, sports and culture.
- SVM Rank uses a linear kernel with trade-off parameter 100, tuned on a held-out ATB development portion. The Java implementation loads in about 5 seconds and tags roughly 2,000 words per second on a laptop.
- The VerbOrNot feature comes from running Al-Khalil over the unique words of ten years of Aljazeera.net articles (95.4M tokens, 613k unique). Each word is labelled V, possible-V or not-V.
- The best bi-LSTM (features only, Farasa segmentation) beat the SVM Rank clitic setup by just 0.1%.
- With automatic segmentation, the bi-LSTM loses more accuracy than SVM Rank, with a 2.6% drop when using embeddings. Word-level bi-LSTM tagging failed, scoring below 82%.
- The leading error types differ by system:
  - SVM Rank: ADJ→NOUN (26.3%).
  - bi-LSTM: segmentation errors (21.8%).

## Interesting ideas
- When neighbouring tags are still unknown, the SVM Rank tagger sums the conditional probabilities over every tag a neighbour could take, so it never commits early to a single guess and avoids cascading mistakes.
- Tagging is framed as ranking: one feature vector is built per candidate tag, and the learner only has to put the correct candidate on top.
- An external morphological analyzer was run once over a huge raw corpus, and its output was reduced to a coarse three-way feature (only-verb / maybe-verb / never-verb). This adds lexical knowledge cheaply, without retraining anything.
- The paper treats improvements as possibly non-additive: embeddings alone and features alone helped, but their combination hurt, which only a full factorial comparison would reveal.
- Evaluation deliberately uses a lexically diverse, out-of-treebank news set and reports results under both gold and automatic segmentation, so the cost of upstream errors is visible.

## For Mova

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

- **I1** (tokens): The en tagger could use the ambiguity class of adjacent tokens (all their lexicon-allowed tags) as context features instead of greedy previous predictions.
- **I2** (tokens): The morphology/lemma stage could rank only the analyses its lexicon proposes, which keeps every decision explainable as "candidate X beat Y because of features F".
- **I3** (world): WordNet/FrameNet-derived POS ambiguity classes could be precompiled per word form and fed to the tagger as a single compact, inspectable feature.
- **I4** (learn): Training experiments should ablate every combination of feature groups rather than assume that stacking helpful components keeps helping.
- **I5** (learn): Tagger and parser evaluation should include an out-of-domain set and a predicted-tokenization condition next to gold, so that pipeline error propagation can be measured.

*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

- [Hamad bin Khalifa University](https://md.research.elamur.ai/free/en/universities/hamad-bin-khalifa-university/) — University, Doha, Qatar · 144 works · 14,175 citations

📚 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 · [🔤 Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/) · 126 works

## Cite

Cite as: Kareem Darwish, Hamdy Mubarak, Ahmed Abdelali and Mohamed Eldesouki. 2017. Arabic POS Tagging: Don't Abandon Feature Engineering Just Yet. Proceedings of the Third Arabic Natural Language Processing Workshop.

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

```bibtex
@inproceedings{darwish2017arabic,
  title = {{Arabic POS Tagging: Don't Abandon Feature Engineering Just Yet}},
  author = {Kareem Darwish, Hamdy Mubarak, Ahmed Abdelali and Mohamed Eldesouki},
  booktitle = {Proceedings of the Third Arabic Natural Language Processing Workshop},
  year = {2017},
  doi = {10.18653/v1/w17-1316},
  url = {https://doi.org/10.18653/v1/w17-1316}
}
```

- [← previous Universal Dependencies Parsing for Colloquial Singaporean English](https://md.research.elamur.ai/free/en/papers/wang-2017-universal-dependencies-parsing-colloquial-singaporean/)
- [next → Learning attention for historical text normalization by learning to pronounce](https://md.research.elamur.ai/free/en/papers/bollmann-2017-learning-attention-historical-text-normalization/)
- [↑ back to papers of 2017](https://md.research.elamur.ai/free/en/papers/2017/page/4/)

## References

12 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Ahmed Abdelali, Kareem Darwish, Nadir Durrani, and Hamdy Mubarak. 2016. Farasa: A Fast and Furious Peilu Wang, Yao Qian, Frank K Soong, Lei He, and Hai Zhao. 2015. A unified tagging solution: Bidirectional lstm recurrent neural network with word embedding. arXiv preprint arXiv:1511.00215. [↗](https://arxiv.org/abs/1511.00215)
2. Thorsten Joachims. 2006. Training Linear SVMs in Linear Time. KDD ’06, pages 217–226, New York, NY. ACM. [find](https://scholar.google.com/scholar?q=Thorsten+Joachims%2E+2006%2E+Training+Linear+SVMs+in+Linear+Time%2E+KDD+%E2%80%9906%2C+pages+217%E2%80%93226%2C+New+York%2C+NY%2E+ACM%2E)
3. Wang Ling, Tiago Luı́s, Luı́s Marujo, Ramón Fernandez Astudillo, Silvio Amir, Chris Dyer, Alan W Black, and Isabel Trancoso. 2015. [Finding function in form: Compositional character models for open vocabulary word representation](https://md.research.elamur.ai/free/en/papers/ling-2015-finding-function-form-compositional-character/). arXiv preprint arXiv:1508.02096. [↗](https://arxiv.org/abs/1508.02096)
4. Heiga Zen and Haşim Sak. 2015. Unidirectional long short-term memory recurrent neural network with recurrent output layer for low-latency speech synthesis. In 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 4470–4474. IEEE. [find](https://scholar.google.com/scholar?q=Heiga+Zen+and+Has%CC%A7im+Sak%2E+2015%2E+Unidirectional+long+short%2Dterm+memory+recurrent+neural+network+with+recurrent+output+layer+for+low%2Dlatency+speech+synthesis%2E+In+2015+IEEE+International+Conference+on+Ac)
5. Mohamed Maamouri, Ann Bies, Tim Buckwalter, and Wigdan Mekki. 2004. The Penn Arabic Treebank: Building a Large-Scale Annotated Arabic Corpus. In NEMLAR conference on Arabic language resources and tools, volume 27, pages 466–467. [find](https://scholar.google.com/scholar?q=Mohamed+Maamouri%2C+Ann+Bies%2C+Tim+Buckwalter%2C+and+Wigdan+Mekki%2E+2004%2E+The+Penn+Arabic+Treebank%3A+Building+a+Large%2DScale+Annotated+Arabic+Corpus%2E+In+NEMLAR+conference+on+Arabic+language+resources+and+tool)
6. Ronaldo Messina and Jerome Louradour. 2015. Segmentation-free handwritten chinese text recognition with lstm-rnn. In Document Analysis and Recognition (ICDAR), 2015 13th International Conference on, pages 171–175. IEEE. [find](https://scholar.google.com/scholar?q=Ronaldo+Messina+and+Jerome+Louradour%2E+2015%2E+Segmentation%2Dfree+handwritten+chinese+text+recognition+with+lstm%2Drnn%2E+In+Document+Analysis+and+Recognition+%28ICDAR%29%2C+2015+13th+International+Conference+on%2C+p)
7. Tomas Mikolov, Wen-tau Yih, and Geoffrey Zweig. 2013. [Linguistic Regularities in Continuous Space Word Representations](https://md.research.elamur.ai/free/en/papers/mikolov-2013-linguistic-regularities-continuous-space-word/). In Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT ’13, pages 746–751, Atlanta, GA, USA.
8. Joakim Nivre, Željko Agić, Lars Ahrenberg, and et al. 2015. Universal dependencies 1.2. LIN-DAT/CLARIN digital library at the Institute of Formal and Applied Linguistics, Charles University in Prague. [find](https://scholar.google.com/scholar?q=Joakim+Nivre%2C+Z%CC%8Celjko+Agic%CC%81%2C+Lars+Ahrenberg%2C+and+et+al%2E+2015%2E+Universal+dependencies+1%2E2%2E+LIN%2DDAT%2FCLARIN+digital+library+at+the+Institute+of+Formal+and+Applied+Linguistics%2C+Charles+University+in+Pragu)
9. Hamid Palangi, Li Deng, Yelong Shen, Jianfeng Gao, Xiaodong He, Jianshu Chen, Xinying Song, and Rabab Ward. 2016. Deep sentence embedding using long short-term memory networks: Analysis and application to information retrieval. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 24(4):694–707. [↗](https://doi.org/10.1109/taslp.2016.2520371)
10. Arfath Pasha, Mohamed Al-Badrashiny, Mona T Diab, Ahmed El Kholy, Ramy Eskander, Nizar Habash, Manoj Pooleery, Owen Rambow, and Ryan Roth. 2014. [MADAMIRA: A Fast, Comprehensive Tool for Morphological Analysis and Disambiguation of Arabic](https://md.research.elamur.ai/free/en/papers/pasha-2014-madamira-fast-comprehensive-tool-morphological/). LREC 2014, pages 1094–1101. European Language Resources Association (ELRA).
11. Barbara Plank, Anders Søgaard, and Yoav Goldberg. 2016. [Multilingual part-of-speech tagging with bidirectional long short-term memory models and auxiliary loss](https://md.research.elamur.ai/free/en/papers/plank-2016-multilingual-part-speech-tagging-bidirectional/). Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 412–418.
12. Ilya Sutskever, James Martens, and Geoffrey E Hinton. 2011. Generating text with recurrent neural networks. In Proceedings of the 28th International Conference on Machine Learning (ICML-11), pages 1017–1024. [find](https://scholar.google.com/scholar?q=Ilya+Sutskever%2C+James+Martens%2C+and+Geoffrey+E+Hinton%2E+2011%2E+Generating+text+with+recurrent+neural+networks%2E+In+Proceedings+of+the+28th+International+Conference+on+Machine+Learning+%28ICML%2D11%29%2C+pages+101)

*Updated 1 November 2026*