Judith Law

My work is focused on hierarchical modeling, multivariate covariance estimation, vector autoregressive processes, small-area estimation, and Monte Carlo expectation maximization with applications in quality measurement, financial econometrics, risk management and high-dimensional statistical learning.  I modeled a multi-level distribution of three-dimensional arrays using a Bayesian approach assuming a separable covariance structure that is the sum of a set of Kronecker products. The goal was to find patterns in the covariance of the survey participants’ responses across quality measures and years to enhance public reporting and quality monitoring.