Congratulations August 2008 Graduates!

08-15-2008

The Department of Statistics would like to congratulate all of its August 2008 graduates. On August 2, 2008, Purdue University and the Department of Statistics awarded degrees to the following people:

Ph.D. Graduates - (Advisor) Dissertation Title

Ms. Lingling An - (Rebecca W. Doerge) "Dynamic Clustering of Time Series Gene Expression"
Mr. David Anderson - (William S. Cleveland) "Multifractal Fractional Sum - Difference Models for Internet Traffic"
Mr. Jinguang (Tony) Li - (Michael (Mihails) Levine) "Stepwise Model Building Approach for Mixed-Effects Models with Random Scale Effects"
Ms. Lei Shu - (William S. Cleveland) "An Additive-Interactive Nonlinear Volatility Model, Estimating and Testing"

M.S. Graduates

Mr. Sehan Kim
Ms. Bo Li
Mr. Ilter Saygin
Ms. Yuhong Wu
Mr. Kai Yang

Graduate Certificates

Mr. Robert Harmon

Undergraduates in Statistics

Mr. Evan Hadiwidjaja
Mr. Matthew Huhn

Actuarial Science Majors

Mr. Evan Hadiwidjaja

So far for the year 2008, the Department of Statistics has graduated 6 Ph.D. students, 16 M.S. students, 5 Graduate Certificate students, 27 Undergraduate Statistics students and 23 Actuarial Science students.

We would like to highlight the achievements of one of the Department's August 2008 graduates. Dr. David Anderson, whose advisor was Professor Cleveland, carried out research in two different areas: statistical modeling and mathematical analysis of Internet traffic; and statistical analysis and modeling for syndromic disease surveillance. Both projects required new developments in time series analysis.

Syndromic disease surveillance is the monitoring of health-related variables to provide alerts that can precede confirmatory diagnosis and signal a possible bioterrorist attack or other public health emergency as rapidly as possible. David's work is part of an effort to build the databases and analysis methods for the State of Indiana surveillance system. He developed nonparametric time series models for daily counts of patient emergency room admissions, categorized by the reported symptoms at the time of admission. Data are from 73 emergency rooms across Indiana. The models, which provide much better fits to effects in the data than past work, are now leading to new visualization tools for State public health officials as well as very effective control charts for disease monitoring.

David also worked on the development of multifractal fractional sum-difference (FSD) time-series models that provide the first parsimonious, validated modeling of the nonlinear, long-range dependent statistical properties of packet arrivals on Internet links. The models have a very simple structure resulting from the discovery that a monotone transformation of the sequence of interarrival times results in a Gaussian long-range dependent sequence consisting of two readily-interpretable additive components: white noise, and a 2-aggregate (moving sum of length two) of discrete fractional Brownian motion. David was able to provide simple intuition and formulas for many traffic phenomena that in the past were described only numerically by estimates of autocorrelations, power spectra, wavelets, and variances of traffic aggregates of different lengths. The work is leading to a better understanding of quality of service issues for the Internet, which depend heavily on the traffic statistics.

David will be an Assistant Professor of Mathematics at Xavier University of Louisiana this fall.

Best wishes to all of our graduates. We look forward to hearing from you!