Identification of influential instances in temporal networks

Abstract

Temporal social network exhibits evolution of a network with different scales of dynamics over a period of time. Detection of community along with its strength is a challenging task in these time varying networks. The impact of the incremental changes in the network over time on community strength must be analyzed to capture its behavior. In this paper, we extended the notion of community detection and its strength of the static network to the temporal social network. A methodology is proposed to study the affect of the incremental links on strength of the detected communities within the network. It also includes a method to identify a set of instances during which information flow is effective by observing the strength pattern of the detected communities. It also demonstrates the utilization of the precise temporal information associated with each interaction. Proposed approach is applied on an email network and identified the instances over which the impact of the incremental links is observed.