Datasets:

Modalities:
Text
Formats:
parquet
ArXiv:
Libraries:
Datasets
pandas
License:
vkhylenko's picture
added dataset formation info
e89f235
|
raw
history blame
542 Bytes
metadata
license: openrail++

Dataset formation:

  1. Filtering Ukrainian tweets so that only tweets containing toxic language remain. Source of Ukrainian data: https://github.com/saganoren/ukr-twi-corpus
  2. Non-toxic sentences were obtained from a previous dataset of tweets as well as sentences from news and fiction from UD Ukrainian IU: https://universaldependencies.org/treebanks/uk_iu/index.html
  3. After that, the dataset was split into a train-test-val and all data were balanced both by the toxic/non-toxic criterion and by data source.