基于DQV词表规范的知识图谱质量模型研究

Research on Quality Model of Knowledge Graph Based on Data Quality Vocabulary

[目的/意义]在人工智能领域,知识图谱作为一种新的知识表示方式备受青睐,将其作为数据支撑的应用程序层出不穷。但是当前针对知识图谱的相关研究主要集中在知识图谱构建技术和应用两方面,而知识图谱质量评估仍处于萌芽阶段,特别是缺少与用户质量需求相关的数据质量模型研究。因此构建知识图谱质量模型对于知识图谱质量评估以及开发高质量应用具有重要意义。[方法/过程]文章通过梳理目前知识图谱各应用领域的业务质量需求,将其与质量维度进行映射,并基于W3C的DQV词表规范,遵循最小化本体原则和重用现有词表原则,采用本体栈的方式,构建出一个可扩展、健壮性强的知识图谱质量模型。[结果/结论]文章提出的质量模型提供了完备的、统一的、规范的术语体系来描述知识图谱质量的各个要素,以此来帮助用户检索符合自身业务需求的知识图谱,达到知识发现的目标。

[Purpose/significance] In the field of artificial intelligence, knowledge graphs are favored as a new knowledge representation method, and applications based on knowledge graphs are emerging one after another.However, the current research on knowledge graph mainly focuses on the construction technology and application of knowledge graph, and the evaluation of knowledge graph is still in its infancy, especially the lack of data quality model research related to user quality requirements.Therefore, the construction of knowledge graph quality model is of great significance.[Method/process] In this paper, by combing the business quality requirements of various application fields of knowledge graph, mapping them with quality dimension, and then based on the DQV vocabulary of W3 C,following the principle of minimizing ontology and reusing the existing vocabulary, a scalable and robust quality model of knowledge graph is constructed by using ontology stack.[Result/conclusion] The quality model proposed in this paper provides a complete, unified and standardized terminology system to describe the various elements of knowledge graph quality, so as to help users retrieve the knowledge graph that meets their own business needs and achieve the goal of knowledge discovery.