Logo Lanfrica

KanuriSenti Dataset

Domaine:

natural language processing

Type de record:

dataset
Créateur:
Mai
Éditeur:
KIIYob
Éditeur:
Men
Hôte:avatar
We first leveraged the PanLex lexical database and a Manga Kanuri Dictionary to compile an initial Kanuri wordlist as a foundation for our dataset. We then partnered with native Kanuri speakers from Yobe State University, the University of Maiduguri, and surrounding communities to validate and expand these entries, gather short texts, and ultimately build a comprehensive annotated corpus. Thereafter some frequently occuring emotional words were derived and emotional rating was carried out for the valence, arrousal and domimance of those words as e-anew data.