Engineering Graph Clustering Algorithms

Networks in the sense of objects that are related to each other are ubiquitous. In many areas, groups of objects that are particularly densely connected, so called clusters, are semantically interesting. In this thesis, we investigate two different approaches to partition the vertices of a network into clusters. The first quantifies the goodness of a clustering according to the sparsity of the cuts induced by the clusters, whereas the second is based on the recently proposed measure surprise.