基于共同用户的跨网络分析:社交媒体大数据中的多源问题

Overlapped user-based cross-network analysis:Exploring variety in big social media data

社交媒体大数据是大数据的重要组成部分.与大数据的"4V"特性对应,本文主要讨论社交媒体大数据中的Variety-多源问题.社交媒体的多源主要体现在不同社交媒体网络所关注的异构用户行为信息,理解社交媒体多源现象对于社交媒体分析和社交媒体大数据的深度应用具有重要意义.社交媒体数据具有来源于用户、服务于用户的特点.我们提出从多个社交媒体网络的共同用户入手来进行社交媒体多源分析:(1)跨网络用户建模,整合分散在不同社交媒体网络的行为信息得到完整用户模型,进行个性化服务;(2)多源数据知识关联,以共同用户与多源数据的交互作为桥梁,挖掘多源数据知识关联,服务于社交媒体协同应用.

Social media contributes much to big data. Among the ＂4V＂ characteristics of big data, this article focuses on investigating the ＂variety＂ in big social media data. Social media variety mainly concerns with the heterogeneous user behaviors in differenet social media networks. Understanding into social emdia variety plays important roles in insightful social media analysis and comprehensive social media applications. Social meida is typically generated from user and desinged for user services. We propose to explore social media variety by investigating the overlapped users between different social media networks. Two problems are discussed： （1） cross-network user modeling, where the scattered user behaviors are integrated for complete user modeling and personalized service development; （2） heterogeneous knowledge association, where the overlapped users serve as bridge to mine the cross-network knowledge association and applied in social media collaborative applications.