The Amharic Hate Speech data is collected using the Twitter API spanning from October 1, 2020 - November 30, 2022, considering the socio-political dynamics of Ethiopia in Twitter space. We used WEbAnno tool for data annotation; each tweet is annotated by two native speakers and curated by one more experienced adjudicator to determine the gold labels. A total of 15.1k tweets consisting of three class labels namely: Hate, Offensive and Normal are presented. Read our papers for more details about the dataset (see below).
The dataset is annotated by two annotators and a curator to determine the gold labels.
For more details, You can read our paper
- train/dev/test.csv
This dataset is ready-made for direct experiments and contains tweet_Id, tweet, and labels for the train/dev/test split datasets.
- with_annotators_train/test/dev.csv
This dataset is presented with every detail of annotations by two native speakers and the curator for researchers who want to explore our dataset in more detail.
Exploring Amharic Hate Speech Data Collection and Classification Approaches
@INPROCEEDINGS{ayele-exploring-hate-2023,
author={Ayele, Abinew Ali and Yimam, Seid Muhie and Belay, Tadesse Destaw and Asfaw, Tesfa and Biemann, Chris},
booktitle={Proceedings of the 14th International Conference on RECENT ADVANCES IN NATURAL LANGUAGE PROCESSING (RANLP 2023},
title={Exploring Amharic Hate Speech Data Collection and Classification Approaches},
pages={59--59},
year={2023},
location = {Varna, Bulgaria}
}