faculty:

Jianqing Fan
Harry Hurd
J.S. Marron
Andrew B. Nobel
Gordon Simons
Richard L. Smith

abstracts to view
postscript papers to download 



A note to prospective students: In deciding which graduate school to attend, one of the most important factors will be the research interests of the faculty members, and as a subset, what topics recent PhD students have chosen as dissertation projects. Below is a sampling to give you an idea of what is currently happening at Carolina. Not all faculty have downloadable papers and abstracts yet, but a good overview of their interests can be obtained from their personal homepages. You are encouraged to contact individual faculty members by email if you have any questions or would like specific information.

Jianqing Fan

Recently, Fan has been particularly interested in high-dimensional data analytic modeling, analysis of functional data, mathematical finance and interdisciplinary collaboration. Some recent titles of preprints are: 
Simultaneous confidence bands and hypothesis testing in varying-coefficient models (with Wen-yang Zhang), Jan. 1999

 Goodness-of-fit test for parametric regression models (with Lishan Huang), April 1998

 Statistical inferences for generalized varying-coefficient models (with Zongwu Cai and Run-Ze Li), March 1998 

Functional-coefficient regression models for nonlinear time series (with Zongwu Cai and Qiwei Yao, March 1998 

Geometric understanding of likelihood ratio statistics (with Hui-Nien Hung and Wing-Hiung Wong), March 1998. 

Downloadable postscript files of these and more preprints are found at Preprints; Go here for Abstracts. 

E-mail jfan@stat.unc.edu | homepage

Harry Hurd
Hurd's primary interest is nonstationary random processes with emphasis on periodically correlated processes. This theoretical interests include Fourier theory of correlation; his application-oriented interests are primarily on time-series analysis for nonstationary processes.
Downloadable papers are available from his homepage and include:

Representation and estimation for priodically and almost periodically correlated random processes, in Cyclostationarity i Communications and Signal Processing", W.A.Gardner, ed., IEEE Press, 1993

The Wold isomorphism for cyclostationary sequences (with T. Koski), 1997

Correlation and spectral theory for periodically corerelated fields indexed on Z^Z (with G. Kallianpur and J. Farshidi), 1997.

A complete list of Hurd's publications is available here. 

E-mail hurd@stat.unc.edu | homepage .




J.S. Marron
Marron's theoretical interests are in smoothing methods for curve estimation as a flexible and powerful approach to data analysis - especially useful in situations where a good parametric model is unknown or there is a need for visual model checking. Mathematical analysis, especially a wide array of asymptotics is a frequently used tool; however, computational, numerical and graphical methods are indispensable. Personal application areas include biology, economics, geology, human movement, image analysis, marketing, opthalmology and software engineering.
Go here to see some movies on smoothing and more.

SiZer for exploration of structures in curves (with P. Chaudhuri)

Significance offeatures via SiZer (with P. Chaudhuri)

Interactive local bandwidth choice (with F. Udina)

Assessing bandwidth selectors with visual error criteria

A personal view of smoothing and statistics 

Marron's complete list of downloadable preprints is available here

E-mail marron@stat.unc.edu | homepage.

Andrew B. Nobel
Nobel is interested in density estimation, empirical model selection, adaptive histogram methods, tree-structured clustering and classification, nonparametric pattern recognition and data compression using vector quantization. Recent preprints include:
Limits to classification and regression estimation from ergodic processes

On regression estimation from ergodic samples with additive noise

Adaptive model selection using empirical complexities

Abstracts of these and other papers are available here.

E-mail nobel@stat.unc.edu | homepage.

Gordon Simons
Gordon Simons' interests are in statistical inference and applied probability. He has recently worked on a sequential statistical model for conducting clinical trials and a collection of inferential questions associated with binary symmetric channels. Among his theoretical interests has been an effort to unify some old work of J.L. Doob's concerned with stopping times, and some recent work by Ignatov concerned with the distribution of record values. A current project is writing a monograph with S. Csörgö on the St. Petersburg Paradox.
 
Simons publications, including downloadable preprints are linked here.

"Asymptotics when the number 0f parameters tends to infinity in the Bradley-Terry model for paired comparisons", to appear in Annals of Statistics.

Email simons@stat.unc.edu | homepage.

Richard L. Smith
Richard Smith's research interests include extreme value theory and its applications to climatology, air pollution, insurance and finance; spatial statistics and environmental applications; Bayesian predictive inference; nonlinear time series and chaos.
A complete list of preprints is available here. The following is a sampling. 

"Bootstrap goodness of fit test for the beta binomial model" (with S. Garren and W. Piegorsch).

"Meteorogically dependent trends in urban ozone" (with L-S Huang).

"Bayesian and frequentist approaches to parametric predictive inference".

"Predictive inference, rare events and hierarchical models".

"Disaggregation of rainfall time series via Gibbs sampling" (with V. Granville).

E-mail rls@stat.unc.edu | homepage




 
Return to:
Top of the page
Faculty homepage
Statistics homepage

March 2003