PIIR/ManCAR

Amazon Reviews 2023 (7 Categories, Post-processed)

Overview

This dataset is a curated and post-processed subset of Amazon Reviews 2023. We select 7 product categories and apply a standard preprocessing pipeline widely used in sequential recommendation research. We adopt the official absolute-timestamp split provided by the corpus.

Included Categories

1. 1.CDsandVinyl
2. 2.Video_Games
3. 3.ToysandGames
4. 4.Musical_Instruments
5. 5.GroceryandGourmet_Food
6. 6.ArtsCraftsand_Sewing
7. 7.Office_Products

Post-processing Pipeline

The dataset is processed per category as follows:

1. 1.Positive Sample Selection: We treat user-item interactions with user ratings greater than 3 as positive samples.

1. 1.K-core Filtering: To improve data quality, we remove users with fewer than 10 interactions in CDs and fewer than 5 interactions in the remaining datasets.

Directory Layout (per category)

Each category has its own folder containing:

• —item.csv: Primarily containing the remapped IDs for the items within that specific category.

• —train.csv: A file containing the interaction sequences used for training the model.

• —valid.csv: A dedicated directory containing the validation sequences to tune hyperparameters and prevent overfitting.

• —test.csv: A directory containing the test sequences.

Licensing & Attribution

This dataset is derived from Amazon Reviews 2023. Please refer to the original dataset page for licensing/usage terms and attribution requirements:

• —https://amazon-reviews-2023.github.io/

If you use this processed dataset, please cite the original dataset and clearly state that you used a post-processed subset with the pipeline described above.

Citation

If you use this dataset, please cite:

@article{mancar2026,
  title={ManCAR: Manifold-Constrained Latent Reasoning with Adaptive Test-Time Computation for Sequential Recommendation},
  author={Kun Yang, Yuxuan Zhu, Yazhe Chen, Siyao Zheng, Bangyang Hong, Kangle Wu, Yabo Ni, Anxiang Zeng, Cong Fu, Hui Li},
  journal={arXiv preprint arXiv:2602.20093},
  year={2026}
}