B1790
Title: Sampling from weighted network models when aggregate information is available
Authors: Axel Gandy - Imperial College London (United Kingdom) [presenting]
Abstract: Many networks have weights attached to their edges, such as the number of liabilities in a financial network. One might be interested in sampling from such weighted networks given information about the aggregate amount of in-/out weights of the nodes. This could be because only such partial information about the network is available or because a conditional statistical test should be performed. We discuss several of these settings and present algorithms that can be used to sample from appropriate conditional distributions.
