中文摘要：

复杂网络社区发现具有重要的研究和应用价值．科学合理的社区评价可促使发现隐含的真实社区结构．复杂关联关系使得具有聚团性质的社区可能是非重叠结构，也可能是重叠结构．虽然当前学者们提出许多专门针对非重叠社区、重叠社区的评价方法，但是，在无法预知真实社区拓扑的情况下，采用不同标准对可能出现的多种结果进行评估不具有可比性．所以，急需能同时评价重叠与非重叠社区的统一评价模型，科学合理的评价模型可以辅助发现合适的社区边界及社区尺度，发现隐含的真实聚团子结构．本文从社区聚集度和社区重叠度两个视角出发，提出了一种适用于重叠与非重叠社区的评价模型，不仅能评价出微结构差异引发的社区聚团属性的变化，而且该评价能在一定程度上制衡社区内部聚团性和社区重叠性．从而辅助发现合适尺度、合适边界的重叠或非重叠社区的作用．通过理论分析和各种数据实验证实了本文所提评价模型的合理性和可用性．

英文摘要：

Discovering communities in complex network is important for understanding network topology and predicting its evolution. In real world, individuals always have many attributes and there always are multiplex relationships between them, the underlying com- munity structure is always overlapping. Some algorithms for uncovering overlapping community have been published in recent years. The common difficulty in this research area is how to get proper community size, how to evaluate the quality of overlapping commu- nities. In this paper we propose a measure model for both disjoint and overlapping communities based on community intra-density and inter overlapping rate. When discovering communities, our measure function can help get community with proper granularity to some extend. Comparing with other evaluation formulas, some experiments show our method is reasonable and usability.