disi-unibo-nlp/feast-foodex2-terms
收藏资源简介:
该数据集包含欧洲食品安全局(EFSA)开发的完整FoodEx2分类法——一个用于描述和编码欧洲消费食品的分层分类系统。数据集结构为单一的train分割,每行对应一个FoodEx2术语(约29,900行),列包括术语代码、术语扩展名称、术语范围说明、主父代码和隐式面描述符。分类法以树状层次结构组织食品术语,每个术语有一个父术语,涵盖从广泛食品组到特定食品项目的多个级别。此外,还包括面描述符,用于注释食品的物理状态、脂肪含量或生产方法等属性。数据集用于支持食品分类和检索研究,特别是与FEAST实验相关的可重复性工作。
This dataset contains the complete FoodEx2 taxonomy developed by the European Food Safety Authority (EFSA) — a hierarchical classification system for describing and coding food items consumed in Europe. The dataset structure includes a single train split with approximately 29,900 rows, each representing a FoodEx2 term, with columns such as termCode, termExtendedName, termScopeNote, masterParentCode, and implicitFacets. The taxonomy organizes food terms in a tree hierarchy where each term has one parent, spanning multiple levels from broad food groups to specific food items. It also includes facet descriptors that annotate properties of food like physical state, fat content, or production method. The dataset is intended for research in food classification and retrieval, particularly for reproducibility in FEAST experiments.




