michsethowusu/dyula-sentiments-corpus

This dataset contains sentiment-labeled text data in Dyula for binary sentiment classification (Positive/Negative). Sentiments are extracted and processed from the English meanings of the sentences using DistilBERT for sentiment classification. The dataset is part of a larger collection of African language sentiment analysis resources. Total samples: 36,382 Positive sentiment: 20884 (57.4%)