作　　者： ; ; ; ; (李泓波）;

机构地区： 哈尔滨工程大学计算机科学与技术学院

出　　处： 《哈尔滨工程大学学报》 2012年第6期745-752,共8页

摘　　要： 针对经典网络社区划分方法存在的划分结果难以理解的问题,基于源自物理学中核子场的拓扑势理论,提出针对具有聚类效应的社会网络和复杂网络的社区结点重要度排序算法.在算法中,首先利用NSP方法(network soft partitionbased on topological potential)依据结点在社区中所起的作用将其分为内部结点和边界结点,其次分别对内部结点和边界结点的重要性进行量化并排序,最后将2个排序结果进行拼接以构成最终的排序结果.实验表明,文中算法不但可以解决前述问题,而且具有和快速排序算法同样的时间复杂度. Since the classic methods of network community partitioning have often yielded ambiguous results, an importance-sorting algorithm of community nodes was proposed based on the theory of the topological potential of a nuclear field in the realm of physics. This new algorithm was designed for social networks and complex networks with the clustering effect. It first classified the nodes into inner nodes and boundary nodes according to their roles in the community using an NSP （ network soft partition） based on topological potential. Next it sorted the inner nodes and boundary nodes respectively according to their quantified importance. Then, the results of the two sortings were merged into a final sorting. Experiments show that this proposed algorithm not only solves the problems in the classic methods, but also boasts the same time complexity.

关 键 词： 复杂网络 社会网络 拓扑势 社区结点 重要度排序

领　　域： [自动化与计算机技术] [自动化与计算机技术]