PIIR/ManCAR

---
license: mit
---

# 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. CDs_and_Vinyl
2. Video_Games
3. Toys_and_Games
4. Musical_Instruments
5. Grocery_and_Gourmet_Food
6. Arts_Crafts_and_Sewing
7. Office_Products

## Post-processing Pipeline
The dataset is processed per category as follows:

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

2. 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:

```bibtex
@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}
}