Ego-Community Evolution Tracking in Instant Messaging Networks
Community detection is one of the most topics that are covered by social network analysis researchers. Early works focused mainly on partitioning networks into several global communities before they target communities evolution over time. The main drawback of such approach is the difficulty to obtain the entire network on which we may process a set of algorithms to track evolution. As a result, many researchers focused on detecting and tracking local dynamic communities.
Basically, detecting and tracking local dynamic communities is a process that moves from a set of nodes to build community structure followed by a structure evolution tracking. A particular type of local communities is called “ego-communities”. Unlike local communities, ego-centered ones have the advantage of being able to expand the community according to the neighborhood step of interest node. They allow, among others, to better identify and track the network elements activities.
Existing ego-community detection algorithms do not support directed networks and are designed for static networks. This failure led us to propose a new solution allowing to detect and track ego-community evolution in directed, weighted and dynamic social networks. We present some illustrative examples to explain the working principle of our solution.
Auteur(s) : Ahmed OULD MOHAMED MOCTAR, Idrissa SARR
Pages : 13-22
Année de publication : 2018
Revue : Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering
N° de volume : 249
Type : Article
Mise en ligne par : SARR Idrissa