Statistical network analysis has been a fruitful direction of research in recent years. However, in many settings, the exact edges of such a network may not be available. In such scenarios it may instead be possible to collect aggregate relational data, detailing the number of edges between a node and specific subpopulations present in a network. Existing probabilistic models for network data can then be extended to this setting.
In this talk we will provide an introduction to statistical models for aggregate relational data before considering new results for community detection for data of this form. In particular, we will examine how to perform community detection for large networks along with examining the empirical properties required for community detection in these models.