Hierarchical Methods for Object Data

AOOD Program working group.

This working group is interested in developing hierarchical modeling approaches for object data, including functions, images, and more general structures like shapes and trees.  The goal is to develop inferential methodology motivated by specific applications yielding complex, structured data.  The idea of hierarchical modeling implies flexible, unified models that can simultaneously take into account variability and structure from multiple sources in the data set, within and between objects, and induced by the design or other measured covariates.  Both Bayesian and frequentist approaches will be considered, and discussion of the connections and distinctions among existing Bayesian and frequentist approaches in the literature will be encouraged.