Fixed link

Optimal timescale of community detection in growing networks

Matus Medo, An Zeng, Yi-Cheng Zhang, Manuel S. Mariani

posted on 28 September 2018

pdf (816 views, 616 download, 0 comments)

Many social, economic, and information systems can be represented as networks that grow with time, which makes it challenging to develop and validate methods to analyze their structure. Static methods are limiting as they miss essential information on the system's dynamics. On the other hand, methods based on multi-layer temporal representations of the data lack clear guidelines on how to partition the input time-stamped data into layers. We focus on the popular community detection problem, which aims to simplify the description of a given network by partitioning its nodes into meaningful groups. We use a multi-layer quality function to show, on both synthetic and real datasets, that the temporal duration of the layers that leads to optimal communities is tightly related to the system's intrinsic aging timescale. The use of temporal information leads to drastically different conclusions on the community structure of real networks, which challenges our current understanding of the large-scale organization of growing networks. Our findings point at the importance of understanding the timescales of the dynamical processes that generated the observed networks in order to properly assess their structural patterns.