Abstract: With the wide application of Internet and computer technology, online social networking service becomes one of the best means to keep contacts with old friends and to make new friends. However, there are still many shortages in Social Networking Services(SNS), such as user relationship deterioration and user loss. In order to solve these problems, this paper proposes a SNS website system based on “dormitory”. A module of user relationship mining is embedded to help users to get their potential friends in advance by weight and potential relationship updating. Experimental results show that this system achieves a good recommending accuracy.

Key words: Social Networking Services(SNS), six degrees of separation, small world phenomenon, user relationship mining

摘要： 针对目前网络社交网站存在的交友形式单薄、好友关系淡化、用户流失等问题，设计基于“寝室”组织形式的社交网站系统，嵌入用户挖掘模块，通过权值更新和潜在关系更新算法帮助用户寻找潜在好友，以达到拓展社交圈的目的。实验结果表明，该系统的好友推荐准确度较高。

关键词: 社交网络服务, 六度分隔, 小世界效应, 用户关系挖掘

CLC Number: 

• TP18