面向大规模在线社交网络的社团抽取算法

针对现有的社团分析算法无法在大规模网络上应用的问题,提出一种社团抽取算法,可以高效地分析网络的社团特征.该方法无需事先获取网络的全部拓扑结构,采用网络搜索与社团判定相结合的思路,可有效地抽取结构未知的社交网络上的某个特定社团,从而使分析超大规模网络社团结构成为可能.在仿真数据集上进行实验,分析抽取准确率的影响因素,得出网络平均度越大抽取准确率越高.进一步实验结果表明,社团抽取算法的准确率与现有方法接近,并且执行效率明显高于现有方法,验证了该算法的可行性和有效性.

Community Extraction Algorithm for Large-Scale Online Social Networks

Since the existing community analysis methods cannot be applied in large-scale networks， a community extraction algorithm is proposed. The community structure can be analyzed effectively with the algorithm. The topology of the network is not needed， with the combination of network search and community detection capabilities， the structure of the particular community can be effectively extracted from the social network with unknown topology. The analyzing of the community structure of large scale network is possible with the algorithm. Experiments on simulation data are performed to analyze the influence factor of accuracy， and it is concluded that the accuracy increases with the average degree. Furthermore， it is found that the accuracy of community extraction algorithm is close to existing methods， and the efficiency is much better， the results show the algorithm is feasible and effective.