小说人物关系提取数据集
收藏资源简介:
本研究构建了一个高质量的中国小说人物关系提取数据集,旨在解决现有关系提取方法在处理小说文本中复杂语境和隐含表达方面的挑战。该数据集基于金庸的经典武侠小说《射雕英雄传》,包含100个角色、1,109个对话单元和3,591个关系实例,每个实例都被标注在三个维度上,总计10,773个关系标签。该数据集为未来研究和数字人文研究提供了可靠的数据支持,并有助于自动构建小说中的人物关系网络。
This study constructs a high-quality Chinese fiction character relation extraction dataset, aiming to address the challenges faced by existing relation extraction methods when dealing with complex contexts and implicit expressions in fictional texts. Based on the classic wuxia novel *The Legend of the Condor Heroes* by Jin Yong, this dataset contains 100 characters, 1,109 dialogue units, and 3,591 relation instances. Each instance is annotated across three dimensions, totaling 10,773 relation labels. This dataset provides reliable data support for future research and digital humanities studies, and facilitates the automatic construction of character relationship networks in fiction.
NCRE-dataset 概述
数据集基本信息
- 名称: NCRE-dataset
- 来源: 论文《Dialogue-Based Multi-Dimensional Relationship Extraction from Novels》(NLPCC2025)
数据集描述
- 用途: 用于从小说中提取基于对话的多维关系
- 类型: 语料库
相关论文
- 标题: Dialogue-Based Multi-Dimensional Relationship Extraction from Novels
- 会议: NLPCC2025




