中文摘要：

移动网络社区发现是面向特定移动用户推广移动网络服务的有效方法.利用实际的移动通信数据构建移动网络模型,提出基于用户网络影响的移动社区发现算法.该算法采用信号传递思想将移动用户对网络的影响转化为欧式空间向量关系,利用欧式距离计算用户相似度,基于高效的仿射传播聚类算法实现了移动通信网络的社区结构检测,同时标示了每个社区的核心用户.最后,通过实验验证了算法的有效性并分析了算法的相关参数选择.

英文摘要：

Mobile network community discovery has been introduced as a new efficient way to disseminate mobile Intemet services to a particular group of mobile users. This paper constructs a mobile network model according to the actual data of mobile communication and proposes a mobile community discovery algorithm based on users＇ network effects. Firstly, the algorithm converts the effect of mo- bile users on the network into the relationship between vectors in Euclid space by signaling transmission. And then Euclidean distance is used to calculate user similarity. Secondly, community structure of the mobile communication network is detected by use of the effi- cient affinity propagation clustering and the corresponding core user of each community is marked. Finally , the algorithm is proved ef- fective and its related parameters choice is analyzed in the experiments.