Scholarlink

<p>Knowledge graphs (KGs) have emerged as a revolutionary advancement in knowledge representation in various domains, providing a powerful framework for understanding and interpreting complex real-world entities. As KGs become increasingly domain-specific, generic and open, their utility is evident, but concerns about quality and robustness persist, requiring comprehensive evaluation methods. This paper presents a careful survey and analysis of KG evaluation methods used in different domains and applications. We examine the range of metrics used to evaluate KG quality, covering areas such as link prediction, triple classification, end-to-end reasoning, and more, and discuss their merits and limitations. Challenges associated with KG evaluation will also be explored. Overall, this survey aims to provide a deep understanding of KG evaluation, offering valuable insights for researchers, practitioners, and developers in this dynamic field.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A detailed analysis into evaluation metrics for knowledge graph evaluation

  • 期刊论文

  • 发表日期：

    2025年8月1日

查看全文
  • Hashmy Hassan,
  • Sudheep Elayidom,
  • M. R. Irshad,
  • Christophe Chesneau

摘要

Knowledge graphs (KGs) have emerged as a revolutionary advancement in knowledge representation in various domains, providing a powerful framework for understanding and interpreting complex real-world entities. As KGs become increasingly domain-specific, generic and open, their utility is evident, but concerns about quality and robustness persist, requiring comprehensive evaluation methods. This paper presents a careful survey and analysis of KG evaluation methods used in different domains and applications. We examine the range of metrics used to evaluate KG quality, covering areas such as link prediction, triple classification, end-to-end reasoning, and more, and discuss their merits and limitations. Challenges associated with KG evaluation will also be explored. Overall, this survey aims to provide a deep understanding of KG evaluation, offering valuable insights for researchers, practitioners, and developers in this dynamic field.