Logo Lanfrica

SSA: A Sesotho Sentiment Analysis dataset

Domaine:

natural language processing

Type de record:

dataset
Créateur:
Mokhosi, RefuoeShikali, CasperQin, ZhiguangQiao, Liu
Éditeur:
Zenodo
Hôte:avatar

The Sesotho Sentiment Analysis (SSA) dataset is a labelled Sentiment Analysis(SA) and Aspect Based Sentiment Analysis(ABSA) dataset for the Sesotho language spoken in Southern Africa. It reviews 4402 sentences for the ABSA task and 2363 sentences SA task in the restaurant domain using Lesotho’s orthography of Sesotho. It uses over 43 aspects words for the ABSA task and annotates restaurant review sentences as either positive, negative or neutral.