HUMVI
收藏资源简介:
HUMVI数据集由数据矿工公司和不安全洞察合作创建,旨在通过分析多语言新闻文章来检测影响人道主义援助的暴力事件。该数据集包含17,497篇文章,涵盖英语、法语和阿拉伯语,每篇文章都被标记为相关性,并根据其影响的人道主义援助部门进行分类,如援助安全、教育、食品安全、健康和保护。数据集的创建过程包括与不安全洞察合作,通过人道主义专家验证标签,确保数据集的可靠性和实用性。该数据集主要应用于人道主义援助领域,旨在帮助决策者更好地理解和应对冲突环境中的人道主义需求。
The HUMVI dataset was developed in collaboration between Data Miner Company and Insecurity Insight, with the core objective of detecting violent incidents impacting humanitarian aid through the analysis of multilingual news articles. This dataset comprises 17,497 articles in English, French and Arabic. Each article is annotated for relevance and categorized based on the humanitarian aid sectors it affects, including aid security, education, food security, health and protection. The dataset creation process included label validation conducted by humanitarian experts in partnership with Insecurity Insight to ensure the dataset's reliability and practical value. Primarily utilized in the humanitarian aid domain, this dataset aims to assist policymakers in better understanding and addressing humanitarian needs in conflict settings.




