Representing Ensembles of Networks for Fuzzy Cluster Analysis: a Case Study

As the statistical analysis of networks finds application in an increasing number of disciplines, novel methodologies are needed to handle such complexity. In particular, cluster analysis is among the most successful and ubiquitous data exploration and characterisation techniques. In this work, we focus on how to represent networks ensembles for fuzzy clustering. We explore three different network representations based on probability distribution, autoencoders and joint embedding. We compare de facto standard fuzzy computational procedures for clustering multiple networks on synthetic data. Finally, we apply this approach to a real-world case study.