AraSenCorpus: The corpus contains more than 4.5 million Arabic tweets tagged with 3 sentiment categories:
- Positive
- Negative
- Neutral
The corpus contains an Arabic text in both modern standard Arabic and dialectical Arabic. It is freely available for research purposes only. Kindly cite the following paper:
@article{al2021arasencorpus,
title={Arasencorpus: A semi-supervised approach for sentiment annotation of a large arabic text corpus},
author={Al-Laith, Ali and Shahbaz, Muhammad and Alaskar, Hind F and Rehmat, Asim},
journal={Applied Sciences},
volume={11},
number={5},
pages={2434},
year={2021},
publisher={Multidisciplinary Digital Publishing Institute}
}