Distributed Time-Based Local Community Detection
Apostolos N. Papadopoulos, Georgios Tzortzidis · 2020
The analysis of large time evolving networks poses many challenges due to the immense amount of information available which implies that discovering smaller substructures allows further visualization and analysis that would otherwise be infeasible to compute. Nevertheless, most real-world graphs are constantly changing. Thus their study becomes more challenging. This work addresses the topic of local community detection in dynamic networks, a problem that has recently drawn significant research interest, using PHASR algorithm modified in a way to fit distributed processing standards. In time evolving networks, the challenge is not only to extract the nodes that form the community, but also to find the time interval that spans it’s existence. Some real world cases of local community detection are detecting communities in gene networks to study diseases, in social networks to examine the evolution of friendships and in terrorist networks to reveal terrorist groups.