A0946
Title: Errorfully observed Markov models for evolving networks
Authors: Peter MacDonald - University of Waterloo (Canada) [presenting]
Eric Kolaczyk - McGill University (Canada)
Abstract: A class of continuous-time Markov chain models is considered for binary network data, which evolves over time on an aligned set of nodes. Continuous and discrete-time observation schemes are investigated, where under discrete observation, inference is based on partial observation of the underlying continuous-time process. The statistical properties of some commonly used dynamic network summaries (edge density, number of grown or dissolved edges) are studied towards estimation and inference.
