MidEast-TE
收藏arXiv2024-04-03 更新2024-06-21 收录
下载链接：
https://github.com/yecchen/GDELT-ComplexEvent
下载链接
链接失效反馈官方服务：
资源简介：
MidEast-TE数据集是由新加坡国立大学等机构的研究人员构建，包含约274,795条记录，主要从2015年至2022年间约0.6百万篇新闻文章中提取，重点关注中东地区的国家间合作与冲突事件。数据集创建过程中，研究人员利用预训练的大型语言模型和时间感知聚类技术，自动从新闻文章中提取结构化事件。该数据集适用于时间复杂事件预测，旨在通过分析历史事件，预测未来事件，为灾害预防和早期预警提供支持。
The MidEast-TE dataset was constructed by researchers from institutions including the National University of Singapore. It comprises approximately 274,795 records, primarily extracted from around 600,000 news articles published between 2015 and 2022, with a focus on inter-state cooperation and conflict events in the Middle East region. During the dataset's construction, researchers utilized pre-trained Large Language Models (LLMs) and time-aware clustering techniques to automatically extract structured events from the news articles. This dataset is designed for temporal complex event prediction, aiming to forecast future events through analysis of historical incidents, thereby providing support for disaster prevention and early warning.
提供机构：
新加坡国立大学
创建时间：
2023-12-02
