Skip to main content

Seminar

Academic Careers at Primarily Undergraduate Institutions - Faculty Perspectives from Sciences and Liberal Arts

"Academic Careers at Primarily Undergraduate Institutions - Faculty Perspectives from Sciences and Liberal Arts"

5:00 pm, CP-114B

Professors from local liberal arts institutions will speak about their experiences in this panel discussion.  Faculty speakers include: 

  • Prof. Jennifer Muzyka, Ph.D., Professor of Chemistry, Centre College
  • Prof. Sarah Bray, Ph.D., Associate Professor of Biology at Transylvania University
  • Prof. Saori Hanaki, Assistant Professor of Exercise Science, Pre-Health Committee Advisor, Transylvania University

This event is hosted by the Society of Postdoctoral Scholars.  

Moderators: Prof. Susan Odom (UK chemistry) and Dr. Ellen Crocker (UK forestry)

Date:
-
Location:
Chem-Phys Room 114B
Type of Event (for grouping events):

Penalized Regression via Bayesian Orthant Priors

Dr. Christopher Hans; Associate Professor, Department of Statistics, Ohio State University

Abstract:

Penalized optimization procedures are sometimes described as Bayesian approaches to parameter estimation, prediction and variable selection.  In this talk I examine this connection in the context of the lasso (Tibshirani, 1996) and the elastic net (Zou and Hastie, 2005). Motivated by the proliferation of these (and related) approaches to penalized optimization, I introduce a new class of prior distributions --- the "orthant normal" distribution --- for the regression coefficients in a Bayesian regression model and show that this prior gives rise to the lasso and elastic net point estimates as the posterior mode.  By providing a complete characterization of the prior distribution, I allow for model-based inference that moves beyond exclusive use of the posterior mode, including coherent Bayesian prediction and formal Bayesian model comparison.  In contrast to penalized optimization procedures, where values for the penalty parameters are often selected via cross validation, the Bayesian approach allows for uncertainty about these parameters to be explicitly  included in the model.  I show that the orthant normal distribution has a scale-mixture of  normals representation, providing additional insight into the particular form of shrinkage employed by the elastic net.  Posterior inference is shown to be easily achieved via MCMC methods.   Finally, I generalize the basic orthant normal distribution to allow for a priori dependence in the regression coefficients, yielding a framework for constructing new penalties.  In particular I discuss an approach for incorporating prior information about dependence structure in the covariates that resembles Zellner's g-prior but that also allows for lasso-like shrinkage.

 

Refreshments: 312 MDS building @ 3:30
Date:
-
Location:
MDS 220
Type of Event (for grouping events):

Testing Multinomial Categories with Sequential Sampling: Is a Wheel of Fortune Fair?

Refreshments: 312 MDS building

Abstract: An inverse-type sequential procedure is considered for hypothesis testing of multinomial models. We assume that the probabilities of all k categories (on a wheel of fortune) are at least ε (>0). For developing the procedure Dirichlet Type-II distribution will be primarily used on the basis of a multiple-decision theoretic point of view. A stopping rule is also devised to satisfy a pre-determined P*-condition. Optimal sample sizes will be determined to test various cell configurations in a multinomial model. The developed procedure will be discussed via Monte Carlo study, and for an illustration moderate cell configurations of a wheel of fortune will be considered as well.

Date:
-
Location:
MDS 220
Type of Event (for grouping events):
Subscribe to Seminar