XS-Video
收藏资源简介:
XS-Video数据集是由中国科学院自动化研究所MAIS实验室提出的一个大规模现实世界短视频传播数据集。该数据集收集了来自中国五大平台(抖音、快手、西瓜视频、今日头条、哔哩哔哩)的117720个短视频,包含381926个样本和535个话题,覆盖了从发布后的互动信息,如观看、点赞、分享、收藏、粉丝和评论等。数据集通过跨平台指标对齐方法,对视频的长期传播影响力进行评分,分为0到9级,旨在为短视频传播研究提供全面的互动信息和内容特征。
XS-Video Dataset is a large-scale real-world short video propagation dataset proposed by the MAIS Laboratory of the Institute of Automation, Chinese Academy of Sciences. This dataset collects 117,720 short videos from five major Chinese platforms including Douyin, Kuaishou, Xigua Video, Jinri Toutiao and Bilibili, containing 381,926 samples and 535 topics. It covers post-release interaction information such as views, likes, shares, collections, follower counts and comments. The dataset scores the long-term propagation influence of videos via a cross-platform metric alignment method, with scores ranging from 0 to 9. It aims to provide comprehensive interaction information and content features for short video propagation research.
短视频传播影响力评级数据集概述
数据集基本信息
- 数据集名称:short-video-influence
- 关联研究:Short-video Propagation Influence Rating: A New Real-world Dataset and A New Large Graph Model
数据集内容
- 包含短视频传播影响力评级相关数据
- 提供与大型图模型相关的代码资源
数据用途
- 适用于短视频传播影响力分析研究
- 可用于大型图模型开发与测试




