Multilingual Food Product Entity Matching Evaluation Dataset
- Records
- 6250
- Size
- 0.4GB
- Format
- JSON
- Provider
MOBIUSI INC- Updated
- v1.02026-11-04
This dataset contains food product catalog records for the same items expressed in different languages, transliterations, abbreviations, and localized names, with annotations for product entity relationships. It also includes hard negatives that share similar brands, flavors, or package sizes but refer to distinct products, enabling evaluation of cross-lingual product matching and relationship classification. Typical uses include supermarket catalog deduplication, cross-region catalog alignment, product master data management, and matching model benchmarking.
Overview
Dataset Insights
Preview
Sample Examples
Schema
Technical Specifications
| Field | Type | Description |
|---|---|---|
| file_name | string | File name |
| gold_annotation | object | Product entity relationship and hard negative type labels provided by the dataset creators. |
| evaluation_result | object | The result of comparing the model prediction with the gold annotation. |
| matching_features | object | Comparison features derived from the normalized records for entity matching evaluation. |
| canonicalized_pair | object | Cross-lingual product attributes extracted and normalized from the two original catalog records. |
| predicted_relation | string | The entity relationship predicted by a matching model from the original records and derived features. |
| product_pair_records | object | The two food product catalog records to compare; these form the raw input for each sample. |
| prediction_confidence | number | The model's confidence in its predicted entity relationship, with values ranging from 0 to 1. |
Trust
Compliance Statement
Authorization Type
Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use
Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization
No PII, no real company names, simulated scenarios follow industry standards
Compliance System
Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs
FAQ
Frequently Asked Questions
- What product records does the dataset contain?
- The dataset centers on paired food product catalog records. It covers different language, transliteration, abbreviation, and localized-name forms for the same item, with entity relationship annotations.
- What makes the hard negatives challenging?
- Hard negatives may have similar brands, flavors, or package sizes while representing different products. They help test whether a model can distinguish surface similarity from a true entity match.
- Which capabilities can this dataset evaluate?
- It can support evaluation of cross-lingual product entity matching, catalog record alignment, similar-product discrimination, and retail catalog deduplication.
- How can I access the data and verify its license?
- Refer to the product page and applicable agreement for delivery format, access method, and license scope. Confirm usage rights and restrictions before using the data in commercial or production settings.
Can't find the data you need?
Post a request and let data providers reach out to you.
Get this Dataset
High-quality data, ready for your next AI breakthrough.
Cite this Work
@dataset{Mobiusi2026,
title={Multilingual Food Product Entity Matching Evaluation Dataset},
author={MOBIUSI INC},
year={2026},
url={https://www.mobiusi.com/datasets/311a91ee9918576dc41355cce8838597?dataset_scene_cate_type=9},
urldate={2026-10-28},
keywords={multilingual product matching,food entity matching,cross-lingual product catalog,product entity alignment,retail data evaluation,product catalog deduplication,entity matching benchmark,hard negative evaluation},
version={1.0}
}Using this in research? Please cite us.