多匹配器自动聚合的知识图谱融合系统构建  

Development of infusion system for knowledge graphs based on automatic aggregation of multiple matchers

作　　者：赵晋巍 刘晓鹏 罗威 Zhao Jin-wei;LIU Xiao-peng;LUO Wei(Information Research Center of Military Sciences,Academy of Military Sciences,Beijing 100142,China)

机构地区：[1]军事科学院军事科学信息研究中心

出　　处：《中华医学图书情报杂志》2019年第9期51-57,共7页Chinese Journal of Medical Library and Information Science

摘　　要：总结了知识图谱融合系统的基本框架,并将其作为系统开发的一个基本参考模型。重点研究了匹配器、自适应聚合器等核心组件的设计,实现了多匹配器的自动优化及聚合参数的自适应调节,从而减少了在映射参数设置上的人工干预,提高了融合系统的自动化程度和融合效率。在此基础上,构建了一种基于多匹配器自适应聚合的知识图谱融合系统原型,并进行了初步实验。The basic framework of infusion system for knowledge graphs was summarized and used as a basic reference model for the development of infusion system for knowledge graphs. The key components of matchers and self-adaptive aggregators were designed,which can achieve the automation of multiple matchers and the self-adaptation of aggregated parameters,and reduce the artificial intervention in mapping parameters. A prototype of infusion system for knowledge graphs was developed based on the self-adaptive aggregation of multiple matchers and tested.

关 键 词：知识图谱 知识图谱融合系统 相似度计算 映射 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]