【作者】 于健；

【导师】 戴居丰；

【作者基本信息】 天津大学 ， 通信与信息系统， 2010， 博士

【摘要】 在线社会网络(Online Social Network, OSN)为用户提供了分享、组织、查找资料的重要途径,是一种备受用户青睐的网络交流方式。在线社会网络已成为社会网络(Social Network, SN)、复杂网络以及Internet应用的研究热点。但是在线社会网络中存在着数据集中存储以及用户隐私数据泄漏的问题。本文利用社会学、数据挖掘中的社会网络分析(Social Network Analysis, SNA)方法以及图论、概率论等研究方法,针对在线社会网络存在的上述问题进行深入研究,取得了如下创新性成果:1.建立了一种基于对等计算模式以及网络节点社会属性的对等社会网络(P2P Social Network, P2P SN)。该网络具有两层结构,底层为根据节点社会相似性建立的社区网络,上层是基于Chord的结构化P2P网络。此种层次化的分布式社会网络,变集中式数据管理为分布式数据管理,解决了目前在线社会网络中数据集中存储所带来的弊端。2.提出了一种能够适用于对等社会网络中资源搜索与资源发布的高效的社会感知(social-aware)路由算法。根据节点间的社会距离和社会相似性建立信息传播机制,确定出网络中的关键节点和桥梁节点,并通过关键节点或桥梁节点分别在网内或网间进行资源和消息的转发,从而减少了查询开销,提高了信息传播效率。3.构建了对等社会网络的信誉模型,建立了网络节点信誉评价体系。充分利用节点的社会关系确定节点的信誉值、评价节点的信誉度,并设计出一种每次交易后节点信誉值(含社会信誉及交易信誉)的更新算法。根据节点信誉度确定访问权限以实现授权访问,减少了节点的恶意行为,有效防止了用户隐私数据泄漏。课题结合理论分析、仿真及实验对模型进行了评价和验证。课题研究为优化在线社会网络提供了理论基础和方法支持,因而具有理论意义和实用价值。更多还原

【Abstract】 Online Social Network (OSN) provides the primary method for sharing and constructing and searching information to users. It is becoming increasing popular network communication. Online Social Network has been a research hot spot in social network and complicated network and Internet application. In Online Social Network, data privacy caused by data centralized management is a major challenge. In this project, using Social Network Analysis (SNA) in sociology and Data Mining and Graph theory and probability theory, we do more research on the proposed problem in Online Social Network and obtained the innovative research result with theoretical value and practical significance.1. The distributed P2P Social Network based on peer-to-peer computing model is built up. Using Social Network Analysis (SNA), we establish a hierarchical P2P Social Network model according to the social distance showing social similarity of peers and the centrality metric showing the importance of peers, which changes the centralized data management into distributed management, to resolve the inherent disadvantage of centralized data management.2. We propose highly efficient social-aware routing algorithm on resource searching and publishing in P2P Social Network. Information dissemination mechanisms are set up by using social distance and social similarity among peers. The resource and messages are forwarded through the key peers and bridge peers in the network. The routing algorithm deduces searching overhead and improves information transmission efficiency.3. The trust model is constructed in P2P Social Network, and the trust evaluation mechanism on peer is set up. We exploit social relationship among peers to set the reputation and evaluate the trust, and design trust (including social-trust and transaction-trust) updating algorithm after each transaction. According to peer trust we define the access authority in order to achieve authorized access and deduce the malicious behavior, and effectively avoid the leaking of users’data privacy.Through theoretical analysis and simulation evaluation, we will validate the proposed model and schemes. The research will provide a theoretical foundation and method support for the optimization of next-generation social networks, thus with theoretical significance and practical value.更多还原