This repository contains the work carried out by ELYADATA for the submission to the NER shared task of 2023 in the context of the first Arabic NLP conference, ArabicNLP 2023.
There is also the Data_preprocessing folder which contains the re-sampling experiments, the data cleaning scripts,
and also the scripts for data preparation that convert the Wojood text files from an IOB format, to a format compatible
with each model.
Detailed instructions on how to train the models and perform data-preprocessing are available in the readme.md files
of each folder.
If you use this code please cite our paper :
@inproceedings{DBLP:conf/wanlp/LaouirineEB23,
author = {Imen Laouirine and Haroun Elleuch and Fethi Bougares},
title = {{ELYADATA} at WojoodNER Shared Task: Data and Model-centric Approaches for Arabic Flat and Nested {NER}},
booktitle = {Proceedings of ArabicNLP 2023, Singapore (Hybrid), December 7, 2023},
pages = {759--764},
publisher = {Association for Computational Linguistics},
year = {2023},
url = {https://aclanthology.org/2023.arabicnlp-1.84},
}