HUMSET

HUMSET是一个由数据友好空间等机构创建的多语言数据集，专注于人道主义危机响应领域。该数据集包含来自2018至2021年间全球各种人道主义危机的文档，涵盖英语、法语和西班牙语三种语言。数据集通过专家标注，提供了丰富的文本片段及其分类信息，支持信息抽取和多标签分类任务。HUMSET的创建旨在通过提供高质量的训练数据，推动NLP技术在人道主义响应领域的应用，以加速和优化危机响应过程。HUMSET is a multilingual dataset developed by organizations including Data-Friendly Space, focusing on the domain of humanitarian crisis response. This dataset includes documents spanning various global humanitarian crises between 2018 and 2021, covering three languages: English, French, and Spanish. Annotated by domain experts, it provides rich text segments and their associated classification annotations, supporting downstream tasks such as information extraction and multi-label classification. The creation of HUMSET aims to advance the application of NLP technologies in the humanitarian response domain by supplying high-quality training data, so as to accelerate and optimize the humanitarian crisis response workflow.

搜集汇总

背景与挑战

背景概述

HUMSET是一个多语言人道主义响应文档数据集，由专家标注，涵盖2018至2021年全球46个项目中的灾害事件，包含约17K个英语、法语和西班牙语文档，用于信息提取和分类任务，支持多标签标注。