摘要
讨论了邮件社区的划分和邮件社区的性质,提出一种基于社区中心动态调整的邮件社区划分算法ACCD。算法采用基于邮箱通信行为特征的余弦相似度评估邮箱之间的相似性,并通过社区中心动态调整的方法进行邮件社区的划分。在实际数据集上的实验表明,在较长一段时间内,一个较大的局域网内部的邮件网络呈现显著的小世界网络特征,同时也表明了邮件社区划分算法的合理性。
The mail community partition and the property of the mail community were discussed, an algorithm called ACCD was proposed to partition mail communities based on adjusting community center dynamically. A method of cosine measure was adopted to evaluate the similarity between two mailboxes based on the characteristics of mailboxes＇ communication behavior, and mail communities were partitioned by the way of adjusting the community centers dynamically. The experiments carried on real dataset show that the behavior of sending and receiving mails between mailboxes within a large LAN shows itself remarkable characteristics of small world network. Simultaneously, the experiment results also show that the algorithm of mail community partition in the paper is reasonable.
出处
《计算机应用》
CSCD
北大核心
2008年第B06期146-149,共4页
journal of Computer Applications
基金
国家自然科学基金资助项目(60773048)
河南省高等学校信息网络重点学科开放实验室项目
关键词
数据挖掘
社会网络
邮件社区划分
小世界网络
data mining
social network
mail community partition
small world network
