Chinese-Literature-NER-RE-Dataset

Dataset Overview

Dataset Name

• Chinese-Literature-NER-RE-Dataset

Dataset Purpose

• For Named Entity Recognition (NER) and Relation Extraction (RE) on Chinese literary texts.

Dataset Description

• Detailed dataset description is provided in the arXiv paper.

Tag Set

• Entity tags: defines 7 entity types.
• Relation tags: defines 9 relation types.

Annotation Format

Entity Annotation

• T tag: identifies an entity.
  • Id: unique identifier of the entity in the document, starting from 0 and incremented for each new entity.
  • Type: entity type, corresponding to one of the entity tags.
  • Begin Index: starting index of the entity, starting from 0 and incremented per character.
  • End Index: ending index of the entity, starting from 0 and incremented per character.
  • Value: the word representing the identified object.

Relation Annotation

• R tag: identifies a relation.
  • Id: unique identifier of the relation in the document, starting from 0 and incremented for each new relation.
  • Arg1 and Arg2: the two entities involved.
  • Type: relation type, corresponding to one of the relation tags.

Citation Information

• Authors: Jingjing Xu, Ji Wen, Xu Sun, Qi Su
• Title: A Discourse-Level Named Entity Recognition and Relation Extraction Dataset for Chinese Literature Text
• Year: 2017
• Link: arXiv article link