Summer of V’s: Complex networks, real data by Simon Dobson

Abstract and slides from Simon’s talk at the recent Summer of V’s workshop.

Complex networks and their associated processes let us model phenomena and study their behaviour, opening-up the potential for understanding and for control. However, the very fact that processes can be sensitive to the fine structure of the networks over which they run introduces problems if we try to run simulations over “real” networks gathered either in the wild or from public sources, since we will be subject to sampling errors that — while small — could be dramatic. Moreover in many cases there is only one real network to experiment with, making it harder to employ the statistical techniques we can use with simulated networks.