Sign in to view source links and access this dataset
Description
Amazon-C4 is a dataset of complex product search queries built from the Amazon Reviews 2023 dataset. It contains 21,223 test rows with features like query, item ID, user ID, original rating, and original review. The dataset was created by McAuley-Lab and last updated on Hugging Face in April 2024.
Use Cases
Train product search models based on natural language user queries.
Evaluate query understanding systems based on real user-item interactions.
Analyze the relationship between user reviews and search intent based on the linked review text.
Study user behavior in e-commerce based on query, user ID, and item ID features.
Strengths
Dataset contains 21,223 test rows, providing a substantial sample for evaluation.
Features include both structured (item_id, user_id, ori_rating) and unstructured (query, ori_review) data for multi-faceted analysis.
Based on the established Amazon Reviews 2023 dataset, suggesting a connection to a known data source.
Limitations
Description metadata is limited; actual data quality requires manual inspection after download.
Column-level documentation is absent; field semantics must be inferred after download.
Only the 'test' split is shown in the example; the availability of training/validation splits is unknown.
Provenance
Source
McAuley-Lab
Collection Method
Built based on the Amazon Reviews 2023 dataset, with complex contexts created by ChatGPT.
Freshness
Last updated 2024-04-09 04:13:44
License is unknown and should be verified before use.