知识图谱中链接数据质量评价研究综述

A Review on Quality Assessment for Linked Data of Knowledge Graph

链接数据为知识图谱的主要表现形式,但目前链接数据的发布缺少统一标准,导致数据质量参差不齐.本文回顾了链接数据质量评价的相关研究,并根据链接数据的发展过程,将衡量数据质量的维度划分为7个类型,分别描述每个数据质量维度的特性.同时,具体介绍了冗余度、可信性的量化评价方法及波动性对链接数据质量的影响.

Linked Data is the mainly expression form of knowledge graph, however. We have found that RDF datasets on the Web are of varying quality due to the lack of unified publishing principles, which influences the quality and accessibility of Linked Data. This paper divides the data quality dimensions into 7 types, and presents the charac- teristics of each data dimension. In addition, the quantitative calculation of redundancy and trustworthiness is de- scribed in detail. Our goal is to help the relevant researchers have a comprehensive understanding about the existing work and to promote the further development of Linked Data quality assessment.