Dynamic Tabu Search for Non Stationary Social Network Identification Based on Graph Coloring

We introduce a new algorithm for the identification of Non Stationary Social Networks called Dynamic Tabu Search for Social Networks DTS-SN, that can analyze Social Networks by mapping them into a graph solving a Graph Coloring Problem (GCP). To map the Social Network into an unweighted undirected graph, to identify the users of the Social Networks, we construct a graph using the features that compound the Social Networks with a threshold that indicates if a pair of users have a relationship between them or not. We also take into account the dynamic behavior of the non stationary Social Network, where the relations between users change along time, adapting our algorithm in real time to the new structure of the Social Network.