The dataset for "A Multilingual Short Text Classification Method Based on In-Context Learning"

AfriSenti-SemEval is a dataset for the multilingual short text sentiment classification task covering 12 African languages, which includes African languages from different language families with sample sizes ranging from 1,261 to 22,152 for each language. It features cross-language-family diversity, resource imbalance and short text sparsity, and thus demonstrates strong representativeness in the research on multilingual short text sentiment classification.