基于文献内聚度的非相关文献知识发现排.pdfVIP

情报分析与研究 情报分析与研究 基于文献内聚度的非相关文献知识发现排 序方法研究 1 2 张云秋 　冷伏海 1 (吉林大学公共卫生学院　长春 13002 1) 2 (中国科学院国家科学图书馆 　北京 100 190) 【摘要 】在对现有非相关文献知识发现中间集排序方法进行分析的基础上 , 以共现理论为基础 , 以主题关联度为 着眼点 ,提出基于文献内聚度加权的 B 排序方法 。并以 Swan son 的早期发现之一为基础 ,考察经文献内聚度加权 和逆文献频率加权两种方法排序筛选后 B 的范围以及 目标关联词和 目标关联对的出现情况 , 以此作为评价其对 B 影响的依据 。结果表明基于文献内聚度加权法能显著提高 B 的质量 ,从而提高发现效率 。 【关键词 】非相关文献知识发现 　中间集 　文献内聚度 　M eSH 【分类号 】G353. 1 Research on the L itera ture Cohesion - ba sed Rank ing M ethod of D isjo in t L itera ture - ba sed D iscovery 1 2 Zhang Yunqiu 　L eng Fuhai 1 ( School of Pub lic H ealth , J ilin Un iversity, Changchun 13002 1, Ch ina) 2 (N ational Science L ibrary, Ch inese A cademy of Science s, B eij ing 100 190 , Ch ina) 【Ab stract 】B ased on the analysis of the existing rank ing m ethod s of B collection of disjoint literature - based discovery, th is p ap er p ropo se s literature cohe sion - ba sed rank ing m ethod according to the co - occu rrence theory and the subject rel ative degree. Then , an exp erim ent is conducted comp aring w ith one of Swan son ’s form er discoveries. The size of B and the occu rrence of the target term s and target relation s are exp lored to evaluate the effect on B of the two m ethod s including the literatu re cohe sion - based weight and the inverse docum ent frequency - ba sed weight. The resu lts of the exp erim ent indicate that the literatu re cohesion - based rank ing m ethod can imp rove the quality of B and enhance efficiency of discov ery accordingly. 【Keywords】D isjoint literatu re - based discovery　B collection　L iteratu re cohe sion　M eSH 1　研究背景 　　1986年美国芝加哥大