Network Tight Community Detection

A fundamental technical challenge in the analysis of network data is the automated discovery of communities -groups of nodes that are strongly connected or that share similar features or roles. In this commentary we review progress in the field over the last 20 years. Networks feature in many aspects of our lives, ranging from social networks of personal relationships, both on and off-line, biological networks of interactions between genes, metabolites and neurons, to technological networks such as the Internet, infrastructure networks, and transportation networks. The modern science of networks investigates the structure and function of networks like these, composed of nodes representing the elementary units of the system and the links or edges that connect them 1 . A prominent feature of networks is their community structure -the organization of nodes into groups, where nodes in the same group are strongly connected or share similar features or roles (Fig. 1 ). Algorithmic methods for the detection of communities in networks have found applications in a wide range of disciplines 2 . Early work on the problem goes back to the sociological literature of the 1980s 3 , but the issue came to the wider attention of the physics community 20 years ago following the work of Girvan and Newman 4 . Here we review the advances of the intervening two decades in the detection of community structure in networks.