CHeeSE Dataset
The CHeeSE dataset is a collection of manually annotated Swiss news articles in German. Specifically, the dataset consists of news articles and debate questions pairs, where each pair is annotated with the stance of the article towards the question, the article emotion, and the emotion of each individual paragraph. The data is freely available to the research community and can be used to tackle emotion and stance detection in German. Computational journalism applications, such as recommender systems, could leverage the dataset to balance the exposure of articles' stances.
More details on the dataset can be found in the paper: external pagehttps://aclanthology.org/2021.fever-1.8/
The dataset can be downloaded here: external pagehttps://form.jotform.com/223414125349350