基于社团密度的社团发现算法

发布时间：2018-06-01 14:46

本文选题：社团结构 + 社团发现　；参考：《计算机应用研究》2017年07期

【摘要】：发现社团结构是研究复杂网络的重要前提,目前社团发现算法研究存在两个较为严峻的问题:评价函数单一和经典算法时间复杂度过大并且无法发现小粒度的社团。针对上述问题,提出了一种合理的发现算法评价函数,即社团完整度F。实验证明,与经典的模块度函数Q相比,社团完整度函数F能够更合理地评价社团划分质量且社团完整度函数F的灵敏度高于模块度函数Q;提出了基于社团密度的社团发现算法(BDA算法)。实验证明,该算法不仅可以发现小粒度的社团结构,随着网络节点数和边数的增加,BDA算法在时间复杂度方面也具有明显的优势。尝试将BDA算法应用在科学合作者网络并得到了合理的社团结构。
[Abstract]:The discovery of community structure is an important prerequisite for the study of complex networks. At present, there are two severe problems in the research of community discovery algorithms: the single evaluation function and the large time complexity of classical algorithms and the inability to find small grained communities. In order to solve the above problems, a reasonable evaluation function of discovery algorithm, I. e., community integrity degree, is proposed. Experimental results show that compared with the classical modular degree function Q, The community integrity function F can evaluate the quality of community partition more reasonably, and the sensitivity of community integrity function F is higher than that of module degree function Q, and a community discovery algorithm based on community density is proposed. Experimental results show that the proposed algorithm can not only find the community structure with small granularity, but also has obvious advantages in time complexity with the increase of the number of nodes and edges in the network. This paper attempts to apply the BDA algorithm to the network of scientific collaborators and obtains a reasonable community structure.
【作者单位】：西安建筑科技大学理学院;西安建筑科技大学信息与控制工程学院;
【基金】：国家自然科学基金资助项目(61403298)
【分类号】：O157.5

【相似文献】