面向“人人网”的用户信息采集及拓扑结构测量研究

Measurement Study of Topologies Characteristics for “Renren” Social Networking System

以"人人网"为例,研究社交网站数据采集技术,并对其网络拓扑结构进行详细研究。结果表明:1)"人人网"的节点度分布不同于一般社交网络符合的幂律分布,更倾向于具有指数分布特征,且其度分布具有一定的重尾特性,在小范围内出现了类似小变量饱和现象,并且出现"双峰"现象;2)"人人网"符合小世界特性;3)"人人网"具有同配性,节点度高的节点倾向于与高度节点连接;4)用户状态数、照片数和访客数没有明显的正相关特性。研究成果对于进一步了解社交网络的拓扑结构特征具有重要意义,为后续实现资源监管、跨社交网站的数据挖掘奠定了基础。

In this paper, taking as ＂Renren＂ for example, the social networking site＇s data collection technology is studied. The collected data is used to study ＂Renren＂s topological structure. The results show that, 1） different from general social networks＇ power-law distribution, the node degree distribution of＂Renren＂ tends to follow an exponential distribution; ＂Renren＂s degree distribution has some heavy-tailed feature, and there is a saturation phenomenon of small variables on a small scale; it also presents the ＂double peak＂ phenomenon; 2） ＂Renren＂ has a smaller average shortest path length and a larger clustering coefficient, which means the small world characteristics; 3） ＂Renren＂ shows the assortativity, which means the node with high degrees is inclined to connection to the nodes with high degree; 4） No obvious positive correlation is found in status number, photos number and the visitors number of ＂Renren＂ users. The results are of great significance for the further understanding of the ＂Renren＂ and other social networks＇ topology structure, and they will lay a foundation for resources supervision and cross-social network site＇s data mining.