I. Introduction
Today's real-world networks such as social networks (e.g., Twitter, Facebook, etc.) and collaboration networks are complex and dynamic in nature [1]. Uncovering the dynamics of complex networks (i.e., the networks nodes or edges change over time) is a core research topics over the past few years [2]. Community structure (i.e., group of nodes that are strongly connected internally, but sparsely connected externally) is an important characteristic of the dynamic complex networks [3], [4]. For instance, communities of social networks could be organizational units; communities in the collaboration networks might represent a group of people working together on similar topics. Identifying communities in dynamic networks is a very challenging task. There are several methods that have been introduced for community finding in the last one and half decade, such as modularity optimization [5], clique percolation [6], greedy algorithms [7], label propagation [8], random-walk based [9], and topology based [3] methods. Nearly all these methods are developed for static networks.