Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/unimorph/ckb
https://github.com/unimorph/ckb
Last synced: 28 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/unimorph/ckb
- Owner: unimorph
- Created: 2017-08-03T16:43:52.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2023-07-16T05:17:52.000Z (over 1 year ago)
- Last Synced: 2024-05-01T14:42:47.307Z (8 months ago)
- Size: 2.33 MB
- Stars: 3
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-kurdish - Sorani
README
# Central Kurdish (Sorani - `ckb`)
### *Kurdîy Nawendî* - کوردیی ناوەندی - کوردیی سۆرانی
---
This repository contains inflected forms of words in Central Kurdish. The provided dataset can be used for evaluation and training of morphological analyzers and are included in the [SIGMORPHON shared tasks](https://sigmorphon.github.io/) as well.
The older data on the repository (**before July 2023**) was initially created with paradigms taken from the [Alexina Project](https://almanach.inria.fr/software_and_resources/Alexina-en.html). However, as discussed in the SIGMORPHON 2023 paper entitled [Revisiting and Amending Central Kurdish Data on UniMorph 4.0](https://sinaahmadi.github.io/docs/articles/ahmadi2023sigmorphon.pdf), the previous dataset contained morphological errors which made it impractical for real-world applications.
## Data
This repository provides the following datasets in the two common scripts of Kurdish, the Perso-Arabic and Latin scripts:
| Dataset | # Inflections | Kurdish Perso-Arabic script | Kurdish Latin script |
|-------------------------|-----------------------------|---------------------------------|---------|
| Gold-standard | 997 | [ckb-arab](arab/ckb-arab.tsv) ([train](arab/ckb-arab.train) / [dev](arab/ckb-arab.dev) / [test](arab/ckb-arab.test)) | [ckb-latn](latn/ckb-latn.tsv) ([train](latn/ckb-latn.train) / [dev](latn/ckb-latn.dev) / [test](latn/ckb-latn.test)) |
| Silver-standard | 110,880 | [ckb-arab-large](arab/ckb-arab-large.tsv) ([train](arab/ckb-arab-large.train) / [dev](arab/ckb-arab-large.dev) / [test](arab/ckb-arab-large.test)) | [ckb-latn-large](latn/ckb-latn-large.tsv) ([train](latn/ckb-latn-large.train) / [dev](latn/ckb-latn-large.dev) / [test](latn/ckb-latn-large.test)) |
**For evaluation purposes, we recommend using the gold-standard dataset in either scripts.** You can use the larger dataset (silver-standard) for training purposes as well.
The train-dev-test split has the proportions of 70-10-20 based on the `tsv` files. The inflections in the two scripts are provided in the same order in the files. For instance, line 92 in the gold-standard dataset in the Perso-Arabic script corresponds to the same word form in the Latin script:
Line 92 in [ckb-arab.tsv](arab/ckb-arab.tsv):
`چوون چووە V;PST;IND;ARGNO3S;LGSPEC2`
Line 92 in [ckb-latn.tsv](latn/ckb-latn.tsv):
`çûn çuwe V;PST;IND;ARGNO3S;LGSPEC2`
Given the issues with the previous dataset, we don't recommend using it. However, if you'd need the old UniMorph 2.0 data, you can find it at [UniMorph2.tsv](UniMorph2/UniMorph2.tsv).
## Tags
The gold-standard dataset is created based on a corpus. To provide the context of each word form in the corpus, we additionally provide another dataset called [ckb-context-complete.tsv](ckb-context-complete.tsv). To further clarify language-specific tags (`lgspec`), we use our tags in that dataset.
The UniMorph tags used for Central Kurdish data are as follows:
| Type | Function | Ours | UniMorph |
|--------|-------------------------------------------------------------------|----------------------------|----------------------|
| Affix | Izafe | `[izafe]` | `lgspec1` |
| Affix | postverb adpositions | `[e]` `[ee]` | `lgspec2` |
| Affix | postverb adverbial /ewe/ | `[ewe1]` | `lgspec3` |
| Affix | disc. adpositions | `[da],[ra], [ewe2]` | `lgspec4` |
| Clitic | adverbial clitic | `[ish]` | `lgspec5` |
| Clitic | demonstrative | `[dem]` | `lgspec6` |
| Clitic | copula | `[cop]` | `lgspec7` |
| Clitic | pronominal markers (argument/possessive) on transitive past verbs | `[pm]` | `lgspec8` |
| Clitic | argument markers on noun/adjectives | `[am]` | `lgspec9` |
## Annotators
- Sina Ahmadi (updated dataset)
- Aso Mahmudi (updated dataset)
- Ali Salehi (UniMorph 2.0)
- Géraldine Walther (UniMorph 2.0)
## Paradigm Samples
In Kurdish Perso-Arabic script:
```
لە لە ADP
وڵات وڵاتانی N;PL;LGSPEC1
ماکۆ ماکۆ PROPN
یان یان CONJ
چوارچێوە چوارچێوەی N;LGSPEC1
دهۆک دهۆک PROPN
ناوچە ناوچەی N;LGSPEC1
کاسیاس کاسیاسی PROPN;LGSPEC1
کۆمەڵکوژی کۆمەڵکوژی ADJ
لەئێرە لێرەدا ADP;N;LGSPEC4(DA)
لە+ئەو لەو ADP + ADJ
پەرلەمان پەرلەمانی N;LGSPEC9(3S)
ئەو ئەوەی PRO;LGSPEC6;LGSPEC1
کردن دەکەنەوە V;PRS;IND;ARGNO3P;LGSPEC3
ڕانیە ڕانیە PROPN
سەرۆک سەرۆک N
زانکۆ زانکۆی N;LGSPEC1
هەبوون هەبێت V;PRS;SBJV;ARGAC3S
دان دابوو V;PST;PRF;IND;LGSPEC8(3S)
```
The same forms in the Kurdish Latin script:
```
le le ADP
wiłat wiłatanî N;PL;LGSPEC1
mako mako PROPN
yan yan CONJ
çwarçêwe çwarçêwey N;LGSPEC1
dihok dihok PROPN
nawçe nawçey N;LGSPEC1
kasyas kasyasî PROPN;LGSPEC1
komełkujî komełkujî ADJ
le’êre lêreda ADP;N;LGSPEC4(DA)
le+ew lew ADP + ADJ
perleman perlemanî N;LGSPEC9(3S)
ew ewey PRO;LGSPEC6;LGSPEC1
kirdin dekenewe V;PRS;IND;ARGNO3P;LGSPEC3
řanye řanye PROPN
serok serok N
zanko zankoy N;LGSPEC1
hebûn hebêt V;PRS;SBJV;ARGAC3S
dan dabû V;PST;PRF;IND;LGSPEC8(3S)
```
## Shared Tasks
2021: [SIGMORPHON 2021 Shared Task on Morphological Reinflection: Generalization Across Languages](https://aclanthology.org/2021.sigmorphon-1.25/)
## References
```
@inproceedings{ahmadi2023sigmorphon,
title = "Revisiting and Amending {C}entral {K}urdish Data on {U}ni{M}orph 4.0",
author = "Ahmadi, Sina and Mahmudi, Aso",
booktitle = "Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morphology",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.sigmorphon-1.5",
pages = "38--48"
}
```
```
@inproceedings{pimentel-ryskina-etal-2021-sigmorphon,
title = "SIGMORPHON 2021 Shared Task on Morphological Reinflection: Generalization Across Languages",
author = "Pimentel, Tiago and
Ryskina, Maria and
Mielke, Sabrina J. and
Wu, Shijie and
Chodroff, Eleanor and
Leonard, Brian and
Nicolai, Garrett and
Ghanggo Ate, Yustinus and
Khalifa, Salam and
Habash, Nizar and
El-Khaissi, Charbel and
Goldman, Omer and
Gasser, Michael and
Lane, William and
Coler, Matt and
Oncevay, Arturo and
Montoya Samame, Jaime Rafael and
Silva Villegas, Gema Celeste and
Ek, Adam and
Bernardy, Jean-Philippe and
Shcherbakov, Andrey and
Bayyr-ool, Aziyana and
Sheifer, Karina and
Ganieva, Sofya and
Plugaryov, Matvey and
Klyachko, Elena and
Salehi, Ali and
Krizhanovsky, Andrew and
Krizhanovsky, Natalia and
Vania, Clara and
Ivanova, Sardana and
Salchak, Aelita and
Straughn, Christopher and
Liu, Zoey and
Washington, Jonathan North and
Ataman, Duygu and
Kiera{\'s}, Witold and
Woli{\'n}ski, Marcin and
Suhardijanto, Totok and
Stoehr, Niklas and
Nuriah, Zahroh and
Ratan, Shyam and
Tyers, Francis M. and
Ponti, Edoardo M. and
Aiton, Grant and
Hatcher, Richard J. and
Prud'hommeaux, Emily and
Kumar, Ritesh and
Hulden, Mans and
Barta, Botond and
Lakatos, Dorina and
Szolnok, G{\'a}bor and
{\'A}cs, Judit and
Raj, Mohit and
Yarowsky, David and
Cotterell, Ryan and
Ambridge, Ben and
Vylomova, Ekaterina",
booktitle = "Proceedings of the 18th SIGMORPHON Workshop on Computational Research in Phonetics, Phonology, and Morphology",
month = aug,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.sigmorphon-1.25",
doi = "10.18653/v1/2021.sigmorphon-1.25",
pages = "229--259"
}
```
## License:
[Creative Commons Attribution-ShareAlike 3.0 Unported (CC BY-SA 3.0)](https://creativecommons.org/licenses/by-sa/3.0/)