基于混合聚类的本体分块与映射  

Ontology block matching based on mixed clustering

作　　者：徐德智[1] 杨冠军[1] 陈建二[1] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083

出　　处：《计算机工程与应用》2010年第1期116-118,159,共4页Computer Engineering and Applications

基　　金：国家自然科学基金No.60433020;湖南省自然科学基金No.06JJ50142;湖南省国土资源厅科技计划项目(No.200718)~~

摘　　要：在本体的映射研究中,大规模本体之间的映射一直是研究的难点。当前主要采用分块的思想来处理大本体映射问题。而应用的分块算法只是针对给定分块数的情况。据此,提出基于混合聚类的大本体分块与映射方法(BMC)。该方法首先用语义扩散算法获得结点的语义信息,然后,运用混合聚类算法对本体进行自动分块,最后在各块中进行映射。通过实验结果及分析,表明BMC能取得较好的映射结果。The research on ontology matching between two larger-scale ontologies is still a very challenging work.At present,a solution of block partition is the most popular way to solve the problem,however,the existing algorithm of block partition needs to appoint a fixed number of blocks.To solve this problem,a new block partition algorithm Based on Mixed Clustering （BMC） is proposed.The algorithm firstly uses the semantic diffuse method to obtain node＇s semantic message,and then uses the mixed clustering arithmetic to automatic partition.At last,mappings are discovered in blocks.The results of experiments are shown that BMC can acquire good matching results.

关 键 词：语义扩散 混合聚类 映射 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]