在线网络社区结构发现与演化技术研究

发布时间：2018-04-11 19:06

本文选题：在线网络社区 + 数据采集　；参考：《哈尔滨工程大学》2012年硕士论文

【摘要】：最近几年，随着互联网的快速发展，网络社区用户数量急剧增加，大规模在线网络社区逐渐形成。在线网络社区用户已经成为互联网的重要用户群，在线网络社区应用的重要性已经可以和搜索引擎、即时通讯工具等相提并论。SNS社区、微博等在线网络社区作为新兴的主流互联网应用，已经聚集了大量的用户。随着智能手机等移动终端的发展，网络社区用户数量更是快速增长。网络社区的用户关系是一种复杂网络，复杂网络的社区结构发现一直都是研究热点。通过对现有社区结构发现算法的研究，发现现有的社区结构发现算法关于社区结构或者搜索目标的定义存在过紧或者过松的问题，本文提出一种新的社区结构定义方法；在该方法的基础上，本文还提出了一种基于边介数的快速社区结构发现算法。该算法具有新的社区结构发现策略，并且针对网络中不同的社区结构都能够快速收敛。同时，该算法还提出了关于重叠节点的发现策略，解决了现有很多社区结构发现算法把社区结构发现简单当做图形分割的问题，较好地解决了某些节点同时属于多个社区的情况。网络社区随着时间的推移不断变化，本身具有很强的动态性。社区的节点数量会增加或者减少，节点间关系会变得更加紧密或者疏远；与此同时，网络社区中满足社区结构的节点簇的规模以及它们是否还构成社区都会发生变化。为了预测网络社区在未来的演化状况，需要根据网络社区在过去表现出来的演化特征，建立能够反映社区演化规律的模型。本文针对在线网络社区演化不规则的情况，根据在过去演化过程中各特征的变化率来对未来进行预测，弥补了现有模型对规则特征变化的依赖。模型除了关注网络整体的演化外，还关注了网络中社区的演化规则，确保模型的演化结果跟真实网络社区具有相近的社区结构。在线网络社区结构发现与演化分析基础平台是一个综合性基础分析平台。论文对在线网络社区数据提取方法进行了研究，，针对在线网络社区数据普遍存在限制的情况，提出了一种在线网络社区受限信息的提取方法；同时，在线网络社区的部分数据是动态生成，并不是直接写入静态网页，针对该情况，提出了动态网页数据的提取方案。最后，结合论文中提出的社区结构发现方法、演化分析技术以及社区数据提取方法，实现了在线网络社区结构发现和演化分析基础平台，第五章对该平台的设计和实现方法进行了介绍，平台能够为在线网络社区的结构发现和演化分析提供基础性的支持。
[Abstract]:In recent years, with the rapid development of the Internet, the number of online community users has increased dramatically, and a large scale of online communities have gradually formed.Online network community users have become an important group of Internet users. The importance of online network community applications can be compared with search engines, instant messaging tools, and so on.Weibo and other online communities as a new mainstream Internet applications, has gathered a large number of users.With the development of mobile terminals such as smart phones, the number of network community users is growing rapidly.The user relationship of the network community is a kind of complex network, and the community structure discovery of the complex network is always the research hotspot.Through the research of the existing community structure discovery algorithm, it is found that the existing community structure discovery algorithm has the problem of being too tight or too loose in the definition of community structure or search target. A new community structure definition method is proposed in this paper.On the basis of this method, a fast community structure discovery algorithm based on edge mediums is proposed.The algorithm has a new community structure discovery strategy and can converge rapidly for different community structures in the network.At the same time, the algorithm also puts forward the discovery strategy of overlapping nodes, which solves the problem that many existing community structure discovery algorithms treat community structure discovery simply as graph segmentation.It solves the problem that some nodes belong to multiple communities at the same time.Network community changes with the passage of time, itself has a strong dynamic.The number of nodes in the community will increase or decrease, and the relationship between the nodes will become closer or estranged. At the same time, the size of the nodes that satisfy the community structure in the network community and whether they will also constitute the community will change.In order to predict the evolution of the network community in the future, it is necessary to establish a model that can reflect the evolution law of the network community according to the evolution characteristics of the network community in the past.Aiming at the irregular evolution of the online network community, this paper predicts the future according to the change rate of each feature in the past evolution process, which makes up for the dependence of the existing model on the rule feature change.The model not only pays attention to the evolution of the whole network, but also pays attention to the evolution rules of the community in the network, so as to ensure that the evolution result of the model is similar to that of the real network community.The online network community structure discovery and evolution analysis foundation platform is a comprehensive basic analysis platform.In this paper, the method of online community data extraction is studied, and a method of extracting the limited information of online network community is put forward in view of the limitation of online network community data, at the same time,Part of the data of the online network community is generated dynamically, but not written directly to the static web page. In view of this situation, a scheme of extracting the dynamic web page data is proposed.Finally, combined with the community structure discovery method, evolution analysis technology and community data extraction method proposed in the paper, the online network community structure discovery and evolution analysis platform is realized.The fifth chapter introduces the design and implementation of the platform. The platform can provide the basic support for the structure discovery and evolution analysis of the online network community.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP393.09

【参考文献】