QCRI/CrisisBench-all-lang

CrisisBench数据集是一个用于灾害信息处理和人道主义信息处理的基准测试数据集。该数据集整合了来自多个来源的数据，包括CrisisLex、CrisisNLP、SWDM2013、ISCRAM13、Disaster Response Data (DRD)、Disasters on Social Media (DSM)、CrisisMMD和AIDR。数据集的主要目的是映射类别标签、去除重复数据，并为社区提供基准测试结果。数据集包含多个语言版本，每个JSON对象包含id、event、source、text、lang、lang_conf和class_label等字段。数据集的大小在10K到100K之间，分为train、dev和test三个部分。

The CrisisBench dataset is a benchmark dataset for crisis-related social media data. The dataset includes data from various sources such as CrisisLex, CrisisNLP, SWDM2013, ISCRAM13, Disaster Response Data, Disasters on Social Media, CrisisMMD, and AIDR. The dataset is designed for text classification tasks and includes multiple languages. The data is structured in JSON format with fields such as tweet ID, event name, source, tweet text, language tag, language confidence score, and class label. The dataset is available in two configurations: informativeness and humanitarian, each with train, dev, and test splits. The dataset is licensed under CC BY-NC-SA 4.0.