readme.md

Group-Buying Recommendation Dataset From BeiBei

---

Dataset Files

• data_size.txt
  • Introduction: the count of users and items.
  • Format: <#user>\t<#item>
  • MD5: 9bdec95708f09bed98a077675df794ab
• valid/invalid.txt
  • Introduction: all successful (valid)/failed (invalid) UNDIVIDED records. leave-one-out is applied to valid.txt to generate train_without_invalid/tune/test.txt.
  • Format: (<user_id1>\t<item_id>\t[<user_id2>, ...]) for each line, it's IMPORTANT to note that the first user_id is the initiator's ID and the remaining user_id are the participants' ID (if exists).
  • MD5:
    • 0c22be048a54a7592baf8e1054d4c298 valid.txt
    • 26bb46088f9d6a4dc18aabb94cb832e5 invalid.txt
• train_without_invalid/tune/test.txt
  • Introduction: the positive records for training/tuning (cross-validation)/testing phase.
  • Format: the same as valid/invalid.txt.
  • MD5:
    • 4a881989e3644925fe67629d0a897cec train_without_invalid.txt
    • 5a4807f6329a8b0860e8a47ecaebf4b5 tune.txt
    • 9187a80d67f68830e3daec130979aeaa test.txt
• train.txt
  • Introduction: both successful and failed records for training phase, that is, the combination of ``train_without_invalid.txtandinvalid.txt`.
  • Format: the same as valid/invalid.txt.
  • MD5: 94589c25681d07bf013de7960ef2d8f1
• social_relation.txt
  • Introduction: the social relations among all users.
  • Format: (<user_id1>\t<user_id2>) for each line, meaning that the two users are linked in social networks.
  • MD5: b7b8ad477408193570c80de8073b9a9d

Due to the file size limitation, the negative sample files corresponding to tune/test.txt will be uploaded to GitHub Releases. Each line of the negative sample files is 999 user_id separated by \t and corresponds to tune/test.txt by line number.

Citation

If you want to use our dataset in your research, please cite:

@inproceedings{zhang2021group,
  title={Group-Buying Recommendation for Social E-Commerce},
  author={Zhang, Jun and Gao, Chen and Jin, Depeng and Li, Yong},
  booktitle={2021 IEEE 37th International Conference on Data Engineering (ICDE)},
  year={2021},
  organization={IEEE}
}

Acknowledgement