A collection of multilingual sentiments datasets grouped into 3 classes -- positive, neutral, negative.
Most multilingual sentiment datasets are either 2-class positive or negative, 5-class ratings of products reviews (e.g. Amazon multilingual dataset) or multiple classes of emotions. However, to an average person, sometimes positive, negative and neutral classes suffice and are more straightforward to perceive and annotate. Also, a positive/negative… See the full description on the dataset page:
https://huggingface.co/datasets/tyqiangz/multilingual-sentiments.
