Use of structured, probabilistic models that incorporate multiple layers of uncertainty to describe real-world systems. Topics include multivariate normal distribution, mixture models, nonparametric Bayesian analysis, advanced hierarchical models and latent-variable models, generalized linear models, and advanced topics in linear and nonlinear regression. Three lecture hours per week for one semester. Statistics and Data Sciences 383D and Statistics and Scientific Computation 383D may not both be counted. Prerequisite: Graduate standing; Economics 392M (Topic 19), Statistics and Data Sciences 384 (or Statistics and Scientific Computation 384), or the equivalent; and 383C (or Statistics and Scientific Computation 383C).