---
license: apache-2.0
task_categories:
- text-classification
language:
- en
---

- 44.246 texts in total, 21.493 NOT hateful texts and 22.753 HATE texts
- All duplicate values were removed
- Split using sklearn into 80% train and 20% temporary test (stratified label). Then split the test set using 0.50% test and validation (stratified label)
- Split: 80/10/10
- Train set label distribution: 0 ==> 17.194, 1 ==> 18.202, 35.396 in total
- Validation set label distribution: 0 ==> 2.150, 1 ==> 2.275, 4.425 in total
- Test set label distribution: 0 ==> 2.149, 1 ==> 2.276, 4.425 in total
- Combination of 6 publicly available datasets:
- 1. "Ethos" dataset (Mollas et al., 2022)
- 2. Anatomy of Online Hate: Developing a Taxonomy and Machine Learning Models for Identifying and Classifying Hate in Online News Media (Salminem et al. (2018)
- 3. A Benchmark Dataset for Learning to Intervene in Online Hate Speech (Qian et al., 2019)
- 4. Automated Hate Speech Detection and the Problem of Offensive Language (Davidson, et al., 2017)
- 5. HatEval (Basile et al, 2019), SemEval-2019 Task 5
- 6. "Hate Towards the Political Opponent"(Grimminger et al., 2021)