COMPRISE_Data13_Y-BBCTopics_V1.0

The dataset contains the Yoruba data used to evaluate the performance of different weakly supervised learning techniques for text classification in low-resourced languages. The text was collected from the BBC news website for Yoruba and annotated by two native speakers. The data has seven categories based on the news headlines: Sports, Entertainment, Nigeria, Africa, World, Health, and Politics. It contains 1,908 sentences.