Statistics Discussion Group
Parker Hall 246
Fridays, 2 pm
Each time, a speaker will discuss something selected by the speaker. It is a good opportunity if you are interested in research in Statistics. You can check the full schedule by clicking here
Everyone is welcome.
Speaker: Dr. Guanqun Cao
Title: A Review of Varying Coefficient Regression Models
Abstract: Varying coefficient regression models are known to be very useful tools analyzing the relation between a response and a group of covariates. Their structure and interpretability are similar to those for traditional linear regression model, but they are more flexible because of the infinite dimensionality of the corresponding parameter spaces. The aims of this talk are to give an overview on the recent literature.
Speaker: Ash Abebe
Title: Jaeckel-Hettmansperger-McKean Analysis of the Linear Model
Abstract: I will introduce the Jaeckel-Hettmansperger-McKean (JHM) approach of robust and efficient parameter estimation via minimization of rank-based pseudo-norms. JHM methods have been applied to estimation and testing problems ranging from simple location problems all the way to complex nonlinear regression problems. For location and one-way models, JHM approaches are equivalent to the rank-transformation methods Mann-Whitney-Wilcoxon and Kruskal-Wallis, respectively. However, the rank- transformation method does not extend nicely to models with interaction or linear models in general. My talk will cover JHM estimation in the linear model including a recent extension to hierarchical linear mixed models.
The last part of the talk is based on joint work with Yusuf Bilgic, John Kloke, and Joe McKean.
Speaker: Aaron McAtee
Title: Functional Linear Regression
Abstract: A look at functional linear regression when X(t) is a functional predictor and Y is a real valued response. I will review some methods for choosing the basis functions for the coefficient. A comparison of the various methods will be shown using Canadian weather data.
Speaker: Dr. Peng Zeng
Title: Principal Component Analysis for High-Dimensional Data
Abstract: Principal component analysis (PCA) is a popular procedure for dimension reduction. Targeting the application in high-dimensional data analysis, there are many novel developments on the theory and algorithms of PCA. In this talk, we will review some results in the recent literature.