融合马尔可夫聚类的实体间关系消解方法  

Entity Relation Resolution Method by Integrating Markov Cluster Algorithm

作　　者：常雨骁 庞琳[3] 贾岩涛[1] 林海伦[1,2] 王元卓[1] 刘悦[1] 刘春阳[3] CHANG Yuxiao;PANG Lin;JIA Yantao;LIN Hailun;WANG Yuanzhuo;LIU Yue;LIU Chunyang(Research Center of Web Data Science & Engineering, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China;University of Chinese Academy of Sciences, Beijing 100049, China;National Computer Network Emergency Response Technical Team/Coordination Center of China, Beijing 100029, China)

机构地区：[1]中国科学院计算技术研究所网络数据科学与技术重点实验室,北京100190 [2]中国科学院大学,北京100049 [3]国家计算机网络应急技术处理协调中心,北京100029

出　　处：《计算机科学与探索》2017年第4期511-519,共9页Journal of Frontiers of Computer Science and Technology

基　　金：国家自然科学基金Nos.61173008;61232010;60933005;61402442;61402022;61303244;国家重点基础研究发展计划(973计划)Nos.2013CB329602;2014CB340405;北京市科技新星计划项目No.Z121101002512063;北京市自然科学基金青年基金项目No.4154086;中科院医学影像项目No.KGZD-EW-T03-2;山东省自主创新及成果转化专项No.2014CGZH1103~~

摘　　要：随着面向网络大数据的知识库的不断出现,它们各自都包含海量的实体以及实体间的关系。然而许多有相同含义的关系并没有统一名称,针对这种情况,提出了一种基于马尔可夫聚类(Markov cluster algorithm,MCL)的实体间关系融合方法。该方法首先计算关系间的语义相似度,然后利用关系间的语义相似度作为有边的权重,构建无向图,并利用马尔可夫聚类算法进行聚类。实验表明,该方法相比层次聚类和k-means聚类方法在聚类纯度上有一定提高,并且更加方便使用。Recent years,the development of knowledge bases is very fast.They store large scale of entities and therelations between entities.However,most of the relations which have the same meanings are not in the same form.It is necessary to resolute the relations.For this purpose,this paper proposes an approach based on Markov clusteralgorithm to cluster the relation with same meanings.Firstly,this paper calculates the semantic similarity betweenevery two relations,and then it uses the relation similarity as weighted-edge to build a graph.Finally,this paper runsa Markov cluster algorithm on the graph and gets the result of relation clusters.Experiments show that the proposedapproach has a higher purity than hierarchy cluster and k-means cluster.

关 键 词：马尔可夫聚类 知识库 实体间关系 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]