BERT Models Fine-tuned on Algerian Dialect Sentiment Analysis
These are different BERT models (BERT Arabic models are initialized from AraBERT) fine-tuned on the Algerian Dialect Sentiment Analysis dataset. The dataset contains 50,016 comments from YouTube videos in Algerian dialect. The models are evaluated on the testing set:
| Model Version | No. of Parameters | Training Time | F1-Score | Accuracy |
|---|---|---|---|---|
| LSTM | ~4 M | 3 min | 0.7399 | 0.7445 |
| Bi-LSTM | ~4.3 M | 6 min 35 s | 0.7380 | 0.7437 |
| BERT Base | ~109.5 M | 33 min 20 s | 0.6979 | 0.7500 |
| BERT Large | ~335.1 M | 1 h 50 min | 0.6976 | 0.7484 |
| BERT Arabic Mini | ~11.6 M | 2 min 40 s | 0.7057 | 0.7527 |
| BERT Arabic Medium | ~42.1 M | 11 min 25 s | 0.7521 | 0.7860 |
| BERT Arabic Base | ~110.6 M | 34 min 19 s | 0.7688 | 0.8002 |
| BERT Arabic Large | ~336.7 M | 1 h 53 min | 0.7838 | 0.8174 |
Citation
If you find our work useful, please cite it as follows:
@article{2023,
title={Sentiment Analysis on Algerian Dialect with Transformers},
author={Zakaria Benmounah and Abdennour Boulesnane and Abdeladim Fadheli and Mustapha Khial},
journal={Applied Sciences},
volume={13},
number={20},
pages={11157},
year={2023},
month={Oct},
publisher={MDPI AG},
DOI={10.3390/app132011157},
ISSN={2076-3417},
url={http://dx.doi.org/10.3390/app132011157}
}