Research

Division of Statistics and Data Science faculty is engaged in fundamental and multidisciplinary research in advancing Bayesian statistical modeling by pursuing research in several areas including:

• Bayesian model selection in high-dimensional settings
• Robust and objective Bayesian analyses
• Bayesian categorical data analysis
• Bayesian nonparametrics
• Bayesian computing
• Bayesian application to spatial and spatio-temporal datasets, survey data, biomedical research, and business data

Spatial statistics and uncertainty quantification focus on solving inferential and computational challenges that often arise in the analysis of high-dimensional dependent data, which are prevalent in environmental sciences. Our faculty are advancing knowledge of these fields including:

• Methodology for spatial and spatio-temporal data
• Hierarchical statistical modeling in environmental and climate services 
• Big data problems in uncertainty quantification and spatial statistics
• Uncertainty qualification for computer models 
• Design problems in computer experiments
• Applications in remote-sensing
• Climate change assessments
• Analysis of ecological data

Biological, medical, and health data present numerous challenges such as the prevalence of censored and missing observations, causal inference, and correlated longitudinal data structures, to name a few. Our faculty are advancing knowledge of biostatistics on multiple fronts including:

• Survival analysis; lifetime data analysis with censoring methods
• Statistical causal inference; causal mediational analyses; comparative effectiveness research
• Clinical trials; PK/PD; epidemiology
• Statistical genetics and genomics; bioinformatics
• Multiple imputation for medical research
• Statistical applications in neuroscience

A graphical model is a probabilistic model for which a graph expresses the conditional dependence structure between random variables. They are commonly used in probability theory, statistics, particularly Bayesian statistics, and machine learning. Applications of graphical models include causal inference, information extraction, speech recognition, computer vision, modeling of gene regulatory networks, gene finding and diagnosis of diseases, and for inferring protein structure.

Our faculty conduct various application researches relating to real word marketing or business problems including:

• Statistical modeling for quantitative marketing or economic analysis; econometrics with stochastic frontier models; discrete choice model
• Data editing and imputation 
• Data privacy; statistical disclosure limitation