遇见数据集

Wikidated 1.0

收藏
arXiv2025-09-30 收录
官方服务:

资源简介:

该数据集包含了维基数据(Wikidata)的完整修订历史,通过删除和添加RDF三元组的集合来编码不同版本之间的更改。数据集有两个版本:全局流(global-stream)和实体流(entity-streams),并以JSON Lines格式存储为经过gzip压缩的文本文件。其规模分别为:全局流变体为239吉字节,实体流变体为329吉字节。该数据集的任务是用于分析演变的知识图谱。

This dataset contains the full revision history of Wikidata, where changes between different versions are encoded via sets of added and deleted RDF triples. The dataset has two variants: global-stream and entity-streams, which are stored as gzip-compressed text files in JSON Lines format. The sizes of the two variants are 239 GB for global-stream and 329 GB for entity-streams respectively. This dataset is designed for the task of analyzing evolving knowledge graphs.

提供机构:
Wikidata
二维码
社区交流群
二维码
科研交流群
商业服务