KuaiSAR is a unified search and recommendation dataset containing the genuine user behavior logs collected from the short-video mobile app, Kuaishou (快手), a leading short-video app in China with over 300 million daily active users.
It is the first dataset which records genuine user behaviors, the occurrence of each interaction within either search or recommendation service, and the users' transitions between the two services!
As shown in the following figure, Kuaishou provides both search and recommendation services. The figure illustrates integrated search and recommendation scenarios in Kuaishou app. When watching a video, the user can either scroll up and down to browse different videos with the recommendation service (from middle to left); or tap on the magnifying glass to access the search service (from middle to right).
KuaiSAR has been shared at https://zenodo.org/record/8181109.
OPTION 1. Download via your browser:
You can download the dataset from this link.
Note:
- The 'KuaiSAR_v2.zip' file is for the KuaiSAR dataset.
- The 'KuaiSAR.zip' file is for the KuaiSAR-small dataset.
OPTION 2: Download via the 'wget' command tool:
For the KuaiSAR dataset:
wget https://zenodo.org/record/8181109/files/KuaiSAR_v2.zip
unzip KuaiSAR_v2.zip
For the KuaiSAR-small dataset:
wget https://zenodo.org/record/8181109/files/KuaiSAR.zip
unzip KuaiSAR.zip
If you find it helpful, please cite our paper:
@article{Sun2023KuaiSAR,
title={KuaiSAR: A Unified Search And Recommendation Dataset},
author={Zhongxiang Sun and Zihua Si and Xiaoxue Zang and Dewei Leng and Yanan Niu and Yang Song and Xiao Zhang and Jun Xu},
booktitle={Proceedings of the 32nd ACM International Conference on Information and Knowledge Management},
url = {https://doi.org/10.1145/3583780.3615123},
doi = {10.1145/3583780.3615123},
year={2023},
}
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.