# Data attribution

This noncommercial research prototype derives series-level counts, means, and
SVD item factors from the **Amazon Reviews 2023** Movies & TV category,
released by the McAuley Lab at UC San Diego.

Yupeng Hou, Jiacheng Li, Zhankui He, An Yan, Xiusi Chen, and Julian McAuley.
2024. *Bridging Language and Items for Retrieval and Recommendation.*
https://arxiv.org/abs/2403.03952

Dataset and download documentation: https://amazon-reviews-2023.github.io/

Upstream processing code: https://github.com/hyp1231/AmazonReviews2023

The upstream code repository carries the MIT license (included in
UPSTREAM-LICENSE.txt). That code license is not a claim of ownership of Amazon
reviews or product content. This site publishes no review text, reviewer IDs,
profile information, or product images. It publishes derived model artifacts,
series titles, and an auditable product-to-series mapping. No affiliation with
or endorsement by Amazon, UC San Diego, or the authors is implied.

The raw dataset is not redistributed in this repository. Source download URLs
and SHA-256 digests are documented in the training script/README and model.json.