Logo Lanfrica

cybernaija

Domaine:

natural language processing

Type de record:

dataset
Créateur:
cik
Hôte:
CyberNaija is the first publicly available synthetic cyberbullying corpus in Nigerian Pidgin English (PCM) code-mixed text. It comprises approximately 2,000 samples across four cyberbullying classes -- Harassment, Denigration, Threat, and Exclusion -- plus a Non-Bullying category for classifier training. Samples span three language modes: