Aggregating anonymized Polish Twitter messages categorized into neutral and harmful classes for the PolEval 2019 Cyberbullying Detection task. It provides text data in the `sentence` column and classification labels in the `target` column to facilitate the identification of cyberbullying.
Use Cases
- Train a supervised learning model to detect cyberbullying in Polish using the `sentence` and `target` columns.
- Evaluate the performance of Polish NLP models on anonymized, short-form social media text.
- Conduct toxicity analysis specifically for the Polish language using the provided `target` labels.
Strengths
- Features anonymized Polish-language tweets within the `sentence` field.
- Includes binary classification labels in the `target` column to denote harmful content.
- Derived from the official 2019 PolEval competition task for cyberbullying detection.