Mark Newman
London E1W 1YW, UK
Portland, ME 04101
2nd floor
11th floor
Boston, MA 02115
London E1W 1LP, UK
Talk recording
In most empirical studies of networks, it is assumed that the data we collect accurately reflect the true structure of the network, but in practice this is rarely true: most network data are noisy, containing measurement error, false positives, false negatives, contradictory observations, or missing data. On the other hand the data can also be richly structured, with measurements of different types, repeated observations, annotations, or metadata. This talk will address the problem of making best estimates of network structure from such rich but noisy data, with a variety of example applications in social and biological networks. In the process, we will see that the pattern of errors in network data is far from random and can teach us some intriguing lessons not only about the data but also about the underlying systems they describe.