Network comparison
Abstract
Comparing networks is a key question in network analysis, such as networks of protein-protein interactions in different organisms, or trade networks of different commodities between countries. Often the comparison of interest is based on local structures. A typical approach for comparison of networks is alignment – finding matching nodes and aligning the edges between them as well as possible. This approach has as drawback that usually only a small part of the network can be aligned. Moreover it does not scale well with increasing network size.
Instead we shall use our new comparison network, called Netdis, which is not based on alignment, but rather on counts of small sub-graphs in an ensemble of sub-networks derived from the original networks. With this method, we find that a subsampling approach gives good results even when the comparisons are based only on a small fraction of the networks to be compared.
This is joint work with Waqar Ali, Robert Gaunt and Charlotte Deane.
Background
University Lecturer, Department of Statistics, Oxford, and Fellow at Keble College, Oxford (2000 – present). Senior Research Fellow, King’s College, Cambridge (1998 – 2000). Adjunct Assistant Professor, Department of Mathematics, UCLA, Los Angeles (1996 – 1998). Lecturer, Department of Mathematics, USC, Los Angeles (1994 – 1996). Ph.D. in Mathematics, University of Zurich, Title: A weak law of large numbers for empirical measures via Stein’s method. Advisor: Prof. A.D. Barbour, D.Phil (1994 – present)
Research Interests
Applied Probability, Computational Biology, and Statistics. In particular: Stein’s method, networks, word count statistics