Social network analysis has been studied extensively from variable angles such asdegree distribution analysis,social entity ranking,community extraction,and pattern discovery,etc.In this paper,we consider a person’s social status is highly related with the network struc-ture which he/she locates in,and such impact from network structure to entities’social statuscan be modeled and measured.We extract the dependencies from ordinary relations by analyzingthe link structure.We also propose a “supportiveness”model based on dependency model.Weexploit a supportiveness-based entity ranking scheme,and efficient algorithms are developed tocompute the top-n most supportive entities.Moreover,we extend the supportiveness analysis tocommunity extraction,and develop feasible solutions to identify the most supportive groups ofentities.The empirical study conducted on a large real data set indicates that the supportivenessmeasures are interesting and meaningful,and our methods are effective and efficient in practice.%社交网络分析是近年来的研究热点之一,常见的分析方法包括度分布分析、个体排名、社区发现、模式发现等。文中,作者认为一个人的社会地位与其所在的网络结构具有紧密的联系,而这种网络结构对成员社会地位的影响程度是可以被表示和量化的。文中通过分析社交网络的链接结构,将社交网络中个体与个体间的依赖关系从一般社会关系中抽取出来,提出了一种基于依赖模型的支持力衡量方法,并基于此给出了一种高效的计算最具支持力的节点计算方法。此外,基于上述模型,设计了一种基于依赖关系的支撑结构模型及其计算方法,用于刻画社交网络中特定节点的影响力来源。作者在大规模的真实数据环境下对模型和算法的正确性、效率和伸缩性进行了验证。
展开▼