基于用户查询的多关系群体挖掘改进算法  

An Improved Algorithm of Community Mining from Multi-Relational Network Based on User Inquiry

作　　者：闫娜娜[1] 刘锋[1] 李锡娟[1] 耿波[1] 

机构地区：[1]安徽大学计算机学院,安徽合肥230039

出　　处：《计算机技术与发展》2008年第6期20-22,27,共4页Computer Technology and Development

基　　金：安徽省自然基金项目(070412051);安徽高校省级重点自然科学研究项目(KJ2007A43)

摘　　要：多关系群体挖掘是近年来快速发展的重要的数据挖掘领域之一。传统的群体挖掘方法是假定网络中只有一种关系,并且挖掘结果与用户需求无关。但现实中的社会网络中存在着多种关系。基于用户查询,不同的关系表现出不同的重要性。分析了多关系群体挖掘中关系提取的问题,提出一种新算法对满足用户期望的关系进行最优线性合并。利用获得的合并关系提高群体挖掘的精准性。Community mining in muid- relational network is one of the major fields in data mining which develops fast in recent years. Most of the traditional methods on community mining assume that there is only one kind of relation in the network, and moreover, the mining results are independent of the users＇ needs. However, in reality, there exist multiple relationships in social network, and each of these may play a different role in a particular task. In this paper, analyze relation extraction of community mining from multi - relational network, and propose a new method for learning an optimal linear combination of these relations which can meet the user＇s expection. With the obtained relation, community mining accuracy can be improved.

关 键 词：多关系网络 群体挖掘 关系提取 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]