A0824
Title: Assumption-lean inference for the network-linked data
Authors: Wei Li - Washington University in St. Louis (United States) [presenting]
Robert Lunde - Washington University in St Louis (United States)
Nilanjan Chakraborty - Missouri University of Science and Technology (United States)
Abstract: An assumption-lean inference framework is discussed for regression models built for network-linked data. Two specific network models are explored: the graphon model and the generalized random dot product model, each with different choices of appropriate node-level network statistics. A phase transition phenomenon is discovered in both models. In denser regimes, consistent bootstrap schemes are provided for two important classes of network statistics. In sparser scenarios, a down-sampling inference method is offered that is consistent under mild conditions, albeit with a slightly slower convergence rate.
