Dataset
January 2013
HateSpeech corpus
Over 2,000 posts from the public Polish web, manually annotated for types and degrees of offensive language towards minorities.
HateSpeech corpus in the current version contains over 2000 posts crawled from public Polish web. They represent various types and degrees of offensive language, expressed toward minorities (eg. ethnical, racial). The data were annotated manually.
Bibliography
Marek Troszyński, Aleksander Wawer. Czy komputer rozpozna hejtera? Wykorzystanie uczenia maszynowego (ML) w jakościowej analizie danych. Przegląd Socjologii Jakościowej. 2017. Tom XIII Numer 2. PDF