jjzha/green

该数据集由Green、Maynard和Lin于2022年创建，用于支持职位描述中的实体识别任务。数据集包含句子级别的任务，没有文档分隔符。数据集分为训练集、开发集和测试集，分别包含8669、964和335个样本。标签类型包括通用的BIO标签和更细粒度的标签，如SKILL（技能）、QUALIFICATION（资格）、EXPERIENCE（经验）、OCCUPATION（职业）和DOMAIN（领域），同时还包含词性标注（POS）。数据集来源于TotalJobs（英国）。

原始信息汇总

数据集概述

数据集创建者

• 作者: Green, Thomas; Maynard, Diana; Lin, Chenghua
• 论文标题: Development of a Benchmark Corpus to Support Entity Recognition in Job Descriptions
• 发表会议: Proceedings of the Thirteenth Language Resources and Evaluation Conference
• 发表时间: June 2022
• 出版商: European Language Resources Association

数据集详情

• 许可证: CC-BY-4.0
• 语言: English
• 样本数量:
  • 训练集: 8669 sentences
  • 验证集: 964 sentences
  • 测试集: 335 sentences
• 数据来源: TotalJobs (UK)
• 标签类型:
  • 通用标签: BIO tags with key tags_skill
  • 精细标签:
    • SKILL: Tasks or attributes and abilities.
    • QUALIFICATION: Official certifications.
    • EXPERIENCE: Lengths of time relating to a position or skill.
    • OCCUPATION: Job titles.
    • DOMAIN: Areas of industry.
  • 词性标签: Indicated by pos

示例

{ "idx": 959, "tokens": ["negotiating", "and", "commercial", "skills", "Conscientious", "and", "thorough", "by", "nature"], "tags_skill": ["B-SKILL", "I-SKILL", "I-SKILL", "I-SKILL", "I-SKILL", "O", "B-SKILL", "O", "O"], "pos": ["NN", "CC", "JJ", "NNS", "JJ", "CC", "JJ", "IN", "NN"] }