My main area of research is Network Science, as I work with networks both theoretically as well as practically. What follows is a brief overview of my research interests, with links to my publications in each area.
Analysis of small and large networks
- I run the KONECT project in order to have enough datasets to produce significant research. KONECT is a collection of 230+ networks, and a software library for their generation and analysis.
Spectral and algebraic graph theory
- I have done much work on representing graphs as matrices and using the decomposition of these matrices to get insights on the graphs. My PhD thesis was about the use of graph matrix decompositions in link prediction. The main result of my PhD thesis is the Spectral Evolution Model.
- I showed how split-complex numbers can be used to model relationships on dating websites
- I have contributed to the state of the art in the area of tensor decompositions and joint diagonalisation in network analysis.
Modeling of complex networks
- I have studied processes of network growth based on triangle closing, and the evolution of a network's spectrum.
- I have introduced new measures of inequality, diversity, preferential attachment and bipartivity.
Graphs with special structure
- I started my academic career with work on the Laplacian matrix for signed graphs. Later, I published one of the first online social network dataset with positive and negative edges: the Slashdot Zoo. On a higher level, I found out whether negative links are even necessary in social networks.
- I analysed more exotic matrix decompositions for directed networks, as well as their application to quantum processes.
- I analysed the special properties of bipartite networks, in particular in the context of link prediction.
Machine learning in graphs
- I have developed machine learned methods for performing user role prediction in networks.