Title: Ancestral inference for tree-indexed data
Authors: Anand Vidyashankar - George Mason University (United States) [presenting]
Abstract: Tree-indexed data arise in a variety of applications ranging from cell-kinetics to flow of information in social media. In such problems, it is customary to model the underlying tree as either a discrete-time random tree or a continuous-time random tree. The underlying stochastic processes generating the trees and the indexing data are typically correlated making the data analyses challenging. We describe methods to address the issue of ancestral inference; namely, given the data at some time $t$, how can one construct valid confidence/prediction intervals for parameters of the process at time 0? We develop some new theory to answer such questions and provide few illustrative examples.