POLUSA

POLUSA数据集是由德国弗莱堡大学创建的，包含0.9百万篇涉及政策话题的新闻文章，涵盖2017年1月至2019年8月的时间段。该数据集通过18个新闻出口反映政治光谱，每个出口都根据其政治倾向进行标记。数据集的创建过程包括基础选择、近似重复移除、政策话题选择、政治倾向分配和时间及流行度平衡。POLUSA数据集主要用于研究媒体效应和政治党派性，支持数据密集型深度学习方法的应用。

The POLUSA dataset was created by the University of Freiburg in Germany, consisting of 900,000 news articles focused on policy topics and spanning the period from January 2017 to August 2019. This dataset reflects the political spectrum through 18 news outlets, each labeled according to its political leaning. The dataset creation process includes basic selection, approximate duplicate removal, policy topic selection, political leaning assignment, as well as temporal and popularity balancing. The POLUSA dataset is primarily used for research on media effects and political partisanship, and supports the application of data-intensive deep learning methods.