数显量具族知识图谱的构建及其应用

Construction and application of knowledge graph of digital measuring tools

数显量具种类及其关联信息的全面准确描述有助于对其深刻理解与正确选用,为此探索数显量具族知识图谱的构建与应用问题有重要意义。在数显量具本体的半自动化构建原理的基础上,采用综合算法相关度AR抽取概念,再结合词法模板、 Jaccard系数的多重聚类、 SAO结构的PPMI、 Dice测量强度等算法抽取概念关系,之后使用Protégé工具完成本体构建;提出融合实体关系类别的PRGC实体关系联合抽取模型提取语料中重叠三元组信息。设计原型系统自动化地从网页中抽取与数显量具结构和功能等相关的知识,半监督构建其知识图谱,进一步提供数显量具知识图谱的可视化与知识检索功能,从而提高数显量具知识管理的智能化程度。

The comprehensive and accurate description of the types of digital measuring tools and their associations are helpful for people to deeply understand and select them correctly.For this reason,we explore the construction and application of the knowledge graph of digital measuring tools.First,the semi-automatic construction principle of digital gauge ontology is studied,and concepts are extracted by the comprehensive algorithm AR of domain concept extraction,combined with lexical templates,multiple clustering of Jaccard coefficients,SAO structural PPMI,Dice measurement intensity and other algorithms extract conceptual relationships.And then the Protégétools are used to complete ontology construction.Secondly,a PRGC entity-relation joint extraction model that integrates entity-relationship categories is proposed to extract corpus overlapping triple information.Finally,on this basis,a prototype system is designed to automatically extract knowledge related to the structure and function of digital gauges from web pages,and construct its knowledge graph semi-supervised;and further provide the visualization and knowledge retrieval functions of the knowledge graph of digital gauges.In this way,the intelligent level of knowledge management of digital display measuring tools is improved.