分布式主题地图合并中的TOM算法

为解决分布式环境下主题地图的合并的问题,提出基于主题和事件的合并算法———TOM(Topic&Oc-currence-oriented Merging)算法.TOM算法选择主题和事件的内容作为研究对象,通过主题名称相似性以及事件名称/资源相似性这2个指标,综合判定主题是否可以合并.以IEEE期刊源和EI期刊源作为实验数据源,以2004年度被EI收录的IEEE期刊中研究内容和“knowledge”相关的内容作为考察对象,通过查全率、查准率及其综合指标对TOM算法的质量进行验证,结果表明TOM算法是一种精确度高的优秀的分布式主题地图合并算法.

TOM algorithm in distributed topic maps merging

A TOM(topic & occurrence-oriented merging) algorithm is presented to solve topic maps merging problem in distributed circumstance.The contents of topic and occurrence are selected by TOM as the research objects.Through judging topic name similarity and occurrence data/resource similarity, the conclusion is drawn that whether two topics can be merged.The experimental source data are taken from IEEE and EI.Through evaluating the periodicals published by IEEE in 2004 which contents are "knowledge"-related and can be searchable by EI;the quality of TOM is examined according to the criterions of recall ratio,precision ratio and the compound of them.Finally it is proved that TOM is a kind of precise merging algorithm for distributed topic maps.