Dennis Lin

Penn State University

Dimensional Analysis and Its Applications in Statistics

Dimensional Analysis (DA) is a fundamental method in the engineering and physical sciences for analytically reducing the number of experimental variables prior to the experimentation.  The principle use of dimensional analysis is to reduce from a study of the dimensions of the variables on the form of any possible relationship between those variables.  The method is of great generality.  In this talk, an overview/introduction of DA will be first given.  A basic guideline for applying DA will be proposed, using examples for illustration.  Some initial ideas on using DA for Data Analysis and

Tuesday, October 7, 2014 - 3:30pm

Hokwon Cho

University of Nevada

Sequential Inference for the Risk Ratio and a Measure of Reduction

We propose sequential methods for obtaining approximate confidence limits and optimal sample sizes for the risk ratio (RR) of two independent binomial variates and a measure of reduction (MOR). The procedure is developed based on a modified maximum likelihood estimator (MLE) for the ratio. First-order asymptotic expansions are obtained for large-sample properties of the proposed procedure and we investigate its
finite sample behavior through numerical studies.

Thursday, October 9, 2014 - 3:30pm

Sujit Ghosh

North Carolina State University & SAMSI

A Semiparametric Approach to Source Separation Using Independent Component Analysis

Data processing and source identification using lower dimensional hidden structure plays an essential role in many fields of applications, including image processing, neural networks, genome studies, signal processing and other areas where large datasets are often encountered. Representations of higher dimensional random vector using a lower dimensional vector provide a statistical framework to the identification and separation of the sources.

Thursday, October 30, 2014 - 3:30pm

Bradley Jones


An Efficient Algorithm for Generating Space-filling Designs When the Region of Experimentation is Not Cubic

Latin Hypercube designs (LHD) are in standard use as plans for deterministic computer experiments. However, these designs depend on the ability of the investigator to set each factor independently of all the others. To be specific, the implied design region for an LHD is a square, cube or hypercube. However, there are cases where some parts of such a design region may be inaccessible or even nonsensical. In such cases it is useful to be able to produce a design that is both space-filling while obeying constraints on the design region.

Thursday, September 25, 2014 - 3:30pm

Feifang Hu

George Washington University

Clinical Trials for Personalized Medicine: Design and Statistical Inference

Advances in genetics have allowed scientists to identify genes (biomarkers) that are linked with certain diseases. To translate these great scientific findings into real-world products (personalized medicine) for those who need them, clinical trials play an essential and important role. To develop personalized medicine, we need new designs of clinical trials so that genetics information and other biomarkers can be incorporated in treatment selection.

Thursday, September 11, 2014 - 3:30pm

Bruce Craig

Purdue University

Assessing Inter-Observer Agreement for Multivariate IHC Scoring Data

Immunohistochemical (IHC) staining is widely used in the diagnosis of cancer.

Thursday, September 4, 2014 - 3:30pm

Robert delMas

University of Minnesota

Preliminary Findings from Research on Teaching a Randomization-based Introductory Statistics Course

Preliminary results are presented from an ongoing study of the development of tertiary students’ reasoning in a one-semester college-level statistics course. The modeling and simulation-based course relies on randomization and bootstrap methods for inference. Students in the statistics course learn to use TinkerPlots® to create "just by chance" models that form the basis of simulated distributions of sample statistics in order to draw an inference about an observed effect or difference.

Thursday, August 28, 2014 - 3:30pm

Adrijo Chakraborty

PhD Candidate, University of Georgia Department of Statistics

Hierarchical Bayesian Methods for Survey Sampling and Other Applications

In model-based survey sampling Hierarchical Bayesian (HB) methods have gained immense popularity. One of the major reasons for this popularity remains the convenience in implementation of HB models using MCMC methods even when the models are complex. An inevitable part of this approach is elicitation of the priors for the parameters involved in the model. Authentic expert information can be incorporated by assigning suitable subjective prior distribution to the parameters.

Major Professor(s): 
Dr. Gauri Sankar Datta and Dr. Abhyuday Mandal
Tuesday, June 3, 2014 - 2:00pm
The Cohen Room (230), Statistics Building

Pritam Ranjan

Acadia University

Space-Filling LHDs and Star-Based Multi-Stage Factorial Designs

Space-filling designs are widely used for emulating computer simulators. Over the last three decades, a wide spectrum of Latin hypercube designs (LHDs) has been proposed with different space-filling criteria like minimum correlation among factors, maximin inter-point distance, and orthogonality among the factors. In this talk I will present a new class of space-filling designs. These designs are derived from randomization defining contrast subspaces (RDCSSs) in two-level factorial experiments.

Wednesday, May 21, 2014 - 3:30pm
Statistics Building Cohen Room, 230

Nilanjan Chatterjee

National Cancer Institute

Genetic Architecture of Complex Traits: Implications for Discovery, Prediction and Prevention

Large genome-wide association studies are now consistently pointing towards an extremely polygenic model for complex diseases. Such models may involve thousands of susceptibility markers, each conferring only a modest risk, but collectively they could be explaining substantial variation in disease-risks in populations. Further, a few large studies of gene-environment interactions indicate that genetic and environmental risk-factors may broadly act in a multiplicative fashion on the risk of a number of different cancers and possibly other diseases.

Thursday, September 18, 2014 - 3:30pm


Subscribe to RSS - Colloquium