Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- fr
|
| 4 |
+
viewer: false
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
> [!NOTE]
|
| 8 |
+
> Dataset origin: https://zenodo.org/records/6413804
|
| 9 |
+
|
| 10 |
+
# French verbal frequencies in the Open Subtitle corpus
|
| 11 |
+
|
| 12 |
+
[](http://dx.doi.org/10.5281/zenodo.6412876)
|
| 13 |
+
|
| 14 |
+
This repository provides frequencies for French verbal paradigm cells in
|
| 15 |
+
the [French Open Subtitles Corpus](https://aclanthology.org/L16-1147/).
|
| 16 |
+
|
| 17 |
+
To obtain these frequencies we:
|
| 18 |
+
|
| 19 |
+
- Trained a tagger based on [Flaubert-base-cased](https://huggingface.co/flaubert/flaubert_base_cased) (see [FlauBERT](https://aclanthology.org/2020.lrec-1.302/))
|
| 20 |
+
on the [French_GSD part of the Universal Dependency treebank](https://universaldependencies.org/treebanks/fr_gsd/index.html). In order to improve performances
|
| 21 |
+
on rarer cells, the corpus was augmented with synthetic sentences which we generated using manually written templates and French verbal orthographic data from Demonext.
|
| 22 |
+
- Cleaned and tagged the Open Subtitle Corpus
|
| 23 |
+
- Counted the occurences of each cells for words tagged as VERB or AUX
|
| 24 |
+
- Kept only the predicted feature combinations which were valid French verbal paradigm cells.
|
| 25 |
+
- Converted to the paradigm cell scheme used
|
| 26 |
+
in [Vlexique](http://www.llf.cnrs.fr/fr/flexique-fr.php)
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
The frequencies are extracted from a 401360174 token corpus.
|
| 30 |
+
We found a total of 74148211, from which 415176
|
| 31 |
+
occurences were excluded because of inexistant paradigm cells (prediction errors).
|
| 32 |
+
73733035 remaining occurences are documented.
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
- published under the license: [GPL-3.0](https://opensource.org/licenses/GPL-3.0)
|
| 36 |
+
- version: 1.0.0
|
| 37 |
+
- Contributors:
|
| 38 |
+
- Sacha Beniamine (author)
|
| 39 |
+
- Maximin Coavoux (author)
|
| 40 |
+
- Olivier Bonami (author)
|
| 41 |
+
|
| 42 |
+
Please cite as: Sacha Beniamine, Maximin Coavoux & Olivier Bonami (2022), French verbal frequencies in the Open Subtitle corpus. DOI: 10.5281/zenodo.6412876
|
| 43 |
+
|
| 44 |
+
# Files
|
| 45 |
+
## frequencies (frequencies_subtitles.csv)
|
| 46 |
+
|
| 47 |
+
French paradigm cell frequencies
|
| 48 |
+
|
| 49 |
+
- `cell` (string): paradigm cell, in the vlexique notation.
|
| 50 |
+
- `freq` (integer): total raw frequency for this cell.
|