Seminar Announcements for SPRING 2014

________________________________________

Title: Smoothing Approaches to Partial Differential Equations, With Applications to LIDAR Data
Speaker:  Dr. Raymond J. Carroll, Jill and Stuart A. Harlin ’83 Chair in Statistics Distinguished Professor of Statistics, Texas A&M University
Abstract:   Abstract: Partial differential equation (PDE) models are commonly used to model complex dynamic systems in applied sciences such as biology and finance. The forms of these PDE models are usually proposed by experts based on their prior knowledge and understanding of the dynamic system. Parameters in PDE models often have interesting scientific interpretations, but their values are often unknown, and need to be estimated from the measurements of the dynamic system in the present of measurement errors. Most PDEs used in practice have no analytic solutions, and can only be solved with numerical methods. Currently, methods for estimating PDE parameters require repeatedly solving PDEs numerically with thousands of candidate parameter values, and thus the computational load is high. Here we propose methods to estimate parameters in PDE models, including those with varying coefficients: a parameter cascading method, a Bayesian approach and a kernel-based approach. In the first two methods, the underlying dynamic process modeled with the PDE model is represented via basis function expansion. In linear PDE, simulation studies show that the new methods are comparable, and both outperform other available methods in terms of estimation accuracy. The methods are demonstrated by estimating parameters in a PDE model from LIDAR data.

Date: Fri, Jan, 17,
Time: 3:30pm-4:30pm
Location: Duques Hall, Room 251 (2201 G St NW)
________________________________________

Title: Simulating Realistic Genomic Data With Rare Variants
Speaker:  Dr. Yaji Xu, Department of Statistics, George Washington University
Abstract:   Abstract: Increasing evidence suggests that rare and generally deleterious genetic variants might have a strong impact on disease risks of not only Mendelian disease, but also many common diseases. However, identifying such rare variants remains challenging, and novel statistical methods and bioinformatic software must be developed. Hence, we have to extensively evaluate various methods under reasonable genetic models. Although there are abundant genomic data, they are not most helpful for the evaluation of the methods because the disease mechanism is unknown. Thus, it is imperative that we simulate genomic data that mimic the real data containing rare variants and that enable us to impose a known disease penetrance model. Although resampling simulation methods have shown their advantages in computational efficiency and in preserving important properties such as linkage disequilibrium (LD) and allele frequency, they still have limitations as we demonstrated. We propose an algorithm that combines a regression-based imputation with resampling to simulate genetic data with both rare and common variants. Logistic regression model was employed to fit the relationship between a rare variant and its nearby common variants in the 1000 Genomes Project data and then applied to the real data to fill in one rare variant at a time using the fitted logistic model based on common variants. Individuals then were simulated using the real data with imputed rare variants. We compared our method with existing simulators and demonstrated that our method performed well in retaining the real sample properties, such as LD and minor allele frequency, qualitatively.

Date: Friday, Feb, 07
Time: 11:00-12:00
Location: Phillips Hall, Room 110 (801 22nd Street NW)
________________________________________

Title: (JOINT WITH Decision Sciences ) Incorporating unobserved heterogeneity in Weibull survival models: A Bayesian approach
Speaker:  Dr. Mark Steel, Department of Statistics, Warwick, UK
Abstract:     We propose flexible classes of distributions for survival modelling that naturally deal with both the presence of outlying observations and unobserved heterogeneity. We present the family of Rate Mixtures of Weibull distributions, for which a random effect is introduced through the rate parameter. This family contains i.a. the well-known Lomax distribution and can accommodate flexible hazard functions. Covariates are introduced through an Accelerated Failure Time model and we explicitly take censoring into account. We construct a weakly informative prior that combines the structure of the Jeffreys prior with a proper (informative) prior. This prior is shown to lead to a proper posterior distribution under mild conditions. Bayesian inference is implemented by means of a Metropolis-within-Gibbs algorithm. The mixing structure is exploited in order to provide an outlier detection method. Our methods are illustrated using two real datasets, one concerning bone marrow transplants and another on cerebral palsy.
Date: Thursday, Feb, 13
Time: 4:00-5:00
Location: Phillips Hall, Room 109 (801 22nd Street NW)
________________________________________

Title: Theory and application of large covariance matrix estimation in panel data models
Speaker:  Dr. Yuan Liao, Department of Mathematics, University of Maryland
Abstract:   High dimensional covariance matrix estimation has seen its wide applications in panel data models and factor analysis. While the sparsity assumption on the covariance matrix directly might be restrictive, it is more reasonable to be satisfied when common factors are controlled first. This so-called “conditional sparsity (given factors)” assumption enables us to estimate various covariance matrices with good rate of convergence. Some applications in portfolio allocation are also presented.
Date: Friday, Feb, 28
Time: 11:00-12:00
Location: Phillips Hall, Room 110 (801 22nd Street NW)
________________________________________

Title: TBA
Speaker:  Dr. Nilanjan Chatterjee, National Cancer Institute, Chief of Division of Cancer Epidemiology & Genetics, Biostatistics Branch
Abstract:   Abstract: TBA
Date: Friday, March, 21
Time: 3:30-4:30
Location: Duques Hall, Room 152 (2201 G St NW)
________________________________________

Title: TBA
Speaker:  Dr. Xiaoming Huo, NSF, Program Director Statistics Program (MPS/DMS)
Abstract:   Abstract: TBA
Date: Friday, April, 04
Time: 3:30-4:30
Location: Duques Hall, Room 152 (2201 G St NW)
________________________________________

Title: TBA
Speaker:  Dr. Jacqueline M. Hughes-Oliver, Department of Statistics, George Mason University
Abstract:   Abstract: TBA
Date: Friday, April, 11
Time: 11:00-12:00
Location: Phillips Hall, Room 110 (801 22nd Street NW)
________________________________________

Title: TBA
Speaker:  Dr. Tommy Wright, Chief, Center for Statistical Research and Methodology U.S. Census Bureau
Abstract:   Abstract: TBA
Date: Friday, April, 18
Time: 11:00-12:00
Location: Phillips Hall, Room 110 (801 22nd Street NW)
________________________________________