论文部分内容阅读
文章介绍了社交网络背景下社区的定义以及主要的社区划分评价指标;根据不同的局部优化策略,将基于局部优化的社区发现方法分为局部扩展优化、派系过滤、标签传播、局部边聚类优化4类进行对比分析。基于局部扩展优化的社区发现方法能有效揭示局部社区结构,能提取有意义的局部聚类信息,如层次性和重叠性,对于大规模且动态变化的在线社交网络,在线社区的形成由于依赖局部的交互而表现出更强自治能力,因此局部扩展优化社区发现方法为在线社区挖掘提供了一个非常有效的途径。派系过滤方法由于其严格的社区结构定义能有效发现有结合力的局部社区以及高度重叠社区。标签传播算法在计算复杂度上有着明显的优势,适用于大规模社交网络中的社区挖掘。而基于局部边聚类使社区发现方法能很好地处理网络中的重叠节点。最后,文章对社区发现存在的一些问题和未来的研究做出展望:快速是社区发现方法的一个基本要求和发展趋势;精确性是社区发现技术的一个重要研究方向;综合的分析系统有助于为众多的社区发现技术和方法提供综合、客观的分析和评价;社交网络的动态演化特征给社区发现提出了更高要求和更多挑战。
The article introduces the definition of community in the context of social networks and the main evaluation index of community division. According to different local optimization strategies, the community discovery method based on local optimization is divided into local extension optimization, factional filtering, label propagation, local edge clustering optimization 4 categories for comparative analysis. The community discovery method based on local extension and optimization can effectively reveal the local community structure and extract meaningful local clustering information such as hierarchy and overlap. For the large-scale and dynamic online social network, the formation of online community is dependent on the local Therefore, the local extension and optimization of community discovery methods provide a very effective way for online community mining. Factional filtering can effectively find cohesive local communities and highly overlapping communities because of its strict definition of community structure. Tag propagation algorithm has obvious advantages in computational complexity and is suitable for community mining in large-scale social networks. Based on local edge clustering, community discovery method can well deal with overlapping nodes in the network. Finally, the article forecasts some existing problems and future research on community discovery: fast is a basic requirement and trend of community discovery method; accuracy is an important research direction of community discovery technology; comprehensive analysis system helps Provides a comprehensive and objective analysis and evaluation of numerous community discovery technologies and methods. The dynamic evolution of social networks poses higher requirements and more challenges for community discovery.