知识图谱的候选实体搜索与排序  

Candidate Entity Search and Ranking of Knowledge Graph

作　　者：沈航可 祁志卫 张子辰 岳昆[1] SHEN Hang-Ke;QI Zhi-Wei;ZHANG Zi-Chen;YUE Kun(School of Information Science&Engineering,Yunnan University,Kunming 650500,China)

机构地区：[1]云南大学信息学院,昆明650500

出　　处：《计算机系统应用》2021年第11期46-53,共8页Computer Systems & Applications

基　　金：云南省万人计划“青年拔尖人才”计划(C6193032);云南大学“东陆学者”计划。

摘　　要：根据给定查询实体与知识图谱(Knowledge Graph, KG)中其他实体的相关程度对实体进行排序,是相关实体搜索的重要支撑技术.实体间的相关性不仅体现在KG中,还体现在快速产生的Web文档中.现有的方法主要根据KG来计算实体间的相关度,但KG无法及时地反映真实世界中快速演化的知识,导致计算结果不够客观.因此,本文首先基于TransH模型提出一种候选实体搜索算法,通过分析实体在不同关系超平面中的语义表示来针对不同关系选择候选实体.为了提高候选实体排序的准确性,提出实体无向带权图模型(Entity Undirected Weighted Graph,EUWG),通过量化查询实体与候选实体在Web文档和KG中反映出的相关性,从而准确地对候选实体进行排序.实验结果表明,本文的方法能够在大规模KG中准确地搜索候选实体并对其正确排序.Ranking entities according to the relevance degree between the given entity and other entities in a Knowledge Graph(KG) is critical for related entity search. The relevance between entities is not only reflected in the KG but also the rapidly generated Web documents. In existing methods, the relevance degree is mainly calculated from the KG, which cannot reflect the knowledge rapidly evolving in the real world, and thus effective results cannot be obtained. Therefore,in this study, we first propose an algorithm for searching candidate entities on the basis of the TransH model by analyzing the semantic representation of entities in hyperplanes of different relations. To improve the precision of ranking candidate entities, we propose an Entity Undirected Weighted Graph(EUWG) model by quantifying the relevance between searched and candidate entities reflected in Web documents and KG. Experimental results show that the proposed method can precisely search and rank the candidate entities in the large-scale KG.

关 键 词：知识图谱 相关实体搜索 表示学习 无向带权图 相关度 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]