HQP
收藏资源简介:
HQP数据集是由慕尼黑机器学习中心与慕尼黑大学合作创建的,专门用于检测在线宣传。该数据集包含30,000条英文推文,主要来源于俄罗斯-乌克兰战争期间的社交媒体。数据集通过人工标注和验证生成高质量标签,采用严格的多次批注程序确保标签的准确性。HQP数据集旨在帮助研究社区解决社交媒体中宣传内容检测的问题,特别是在敏感的NLP任务中,如宣传检测,强调了高质量标签的重要性。
The HQP Dataset was co-created by the Munich Center for Machine Learning and Ludwig Maximilian University of Munich (LMU Munich), specifically designed for online propaganda detection. It contains 30,000 English tweets primarily sourced from social media during the Russia-Ukraine War. High-quality labels for the dataset are generated through manual annotation and validation, with a rigorous multi-round annotation procedure implemented to ensure label accuracy. The HQP Dataset aims to assist the research community in addressing the challenge of detecting propaganda content on social media, particularly in sensitive natural language processing (NLP) tasks such as propaganda detection, and underscores the critical importance of high-quality labels.




