基于作品关系扩展的中文同名个人规范记录识别与聚簇研究  

Identification and Clustering of Chinese Same Name Authority Records on Work Relations Extending

作　　者：王瑞云[1] 贾君枝[1] 

机构地区：[1]山西大学经济与管理学院,太原030006

出　　处：《图书情报工作》2017年第5期125-131,共7页Library and Information Service

基　　金：国家社科基金重点项目"基于关联数据的中文名称规范档语义描述及数据聚合研究"(项目编号:15ATQ004)研究成果之一

摘　　要：[目的 /意义]旨在解决中文名称规范联合数据库检索系统CNASS的检索结果集记录量大且杂散的问题,实现其检索服务的关联聚簇功能。[方法 /过程]基于FRBR-LRM框架将个人名称规范记录转换为实体-属性-关系的RDF表示,利用记录内嵌的外部LC记录号重定向到VIAF记录,对原记录的作品关系等属性进行扩展。设计中文同名个人规范记录识别与聚簇算法,充分利用扩展后的作品关系,提高记录识别和聚簇的效率。[结果 /结论]选取300个人名,在CNASS中进行检索,对检索结果集运行算法,统计分析每个检索结果集的聚簇数和最大聚簇内记录数,综合计算聚簇效率指标,验证了本文聚簇算法的有效性。[ Purpose/significance] We manage to deal with the question that the search result set provided by CNASS contains too many records to fulfill clustering and linking for its search service. [ Method/process ] We trans- formed the personal name authority record to RDF representation based on entity-attribute-relations according to the FRBR -LRM frame and extended the work relations of the record with linked LC and VIAF record. We designed identification and clustering algorithms of Chinese same name authority records taking full use of the extended work relations in the re- cord to improve the efficiency of identification and elusteing. [ Result/conclusion] Through the experiment on 300 resuh sets of searching Chinese names on CNASS, the this paper statistically analyzed cluster counters and records within the max-cluster, our identification and clustering method is effective.

关 键 词：名称规范档 实体-属性-关系 作品关系扩展 聚簇 

分 类 号：G254.3[文化科学—图书馆学]