 |
|
|
Galit Shmueli : Personal Web pagePhone: 301-405-9679 Assistant Professor of
Management Science and Statistics |
I've been conducting research in several areas, developing and applying statistical and probabilistic methods. These areas are:
Empirical research of online auctions such as eBay.com has been dominated by researchers from economics and information systems. Together with colleagues and students in the D&IT department and Center for Electronic Markets & Enterprises I have been working on developing a statistical approach and applying state-of-the-art statistical methods for displaying, modeling, and analyzing such data. Bid data are very different from data that are encountered in various fields, and therefore require careful and specialized methods. This wedding of online auction data with a statistical approach is leading to innovations in Statistics as well as in Business. A major focus of ours is the use of functional data analysis (FDA) for analyzing auction dynamics. For papers please visit http://www.rhsmith.umd.edu/ceme/statistics/
I have been involved
in a collaboration between the University of Pittsburgh and Carnegie Mellon
University, entitled the “Computer Based Surveillance Group”, on a
project sponsored by the CDC and the Agency for Healthcare Research and
Quality. The project’s mission is to create a framework for early
detection of Bio-terrorism attacks. The group included researchers from the
areas of epidemiology, public health, computer science, and two
statisticians (Prof. Stephen Fienberg and myself). We used a large
retailer’s database to learn about the potential of using grocery data for
early detection of Bio-terrorism attacks.
The automated detection system that we created is a combination of
techniques from signal processing, machine learning, statistics, and quality
control. This multi-layered detection system uses as input the daily sales
of over-the-counter medication. It monitors the daily sales and compares
each new data point to an expected value. When the two values deviate
significantly, the system signals an alarm. I am currently working on
extensions of the statistical methods in order to track multiple sales
simultaneously using 2D wavelets. For
publications, working papers, and presentations in conferences, please see
my resume.
Runs and Scans
A run is a sequence of consecutive successes in a series of Bernoulli trials. A scan is a “window” of consecutive Bernoulli trials that includes at least a given number of successes. Runs and scans are applied in various fields. Although they are easy to understand and use, the random variables that arise tend to have characteristics (e.g. probability functions, moments) that are complicated for computation.
During my Ph.D. I developed a method for computing exact probabilities for random variables that arise when runs or scans are used. The method is based on Feller’s idea for computing the distribution of the waiting time until the first run (the geometric distribution of order k). In my Ph.D. thesis I generalized Feller’s method using both probability theory and advances in computation power. Since then I have generalized the theory further, and applied it to industrial applications. I wrote two papers, and presented my results in several international conferences as invited and contributed talks. My work is referenced in several books and papers. I am currently working with Prof Louiqa Raschid on applying run theory to the current problem of filtering email from spam, by analyzing sequences of email messages. For publications, working papers, and presentations in conferences, please see my resume.
Web Applications and Industrial Statistics
As many statisticians
have pointed out, there exists an enormous gap between researchers and
practitioners in the field of industrial statistics. In the past few years
several conferences have been organized and associations created
in an attempt to bring both groups together and bridge the gap. A
newly developing approach is to webify theoretical work, thus making it
accessible to users. Since I believe that bridging this gap is essential, I
have dedicated part of my research to creating web applications for widely
used statistical industrial applications. My approach is doubly targeted: To
Webify “good old” existing statistical procedures, tables, and charts
(e.g., the Military Standards) and to webify new theoretical results.
For this reason I created the SQC Online website which is used by many practitioners and academics, and is greatly appreciated. It uses a simple and user-friendly interface to compute probabilities, create graphs, etc., that are then used in industry for various applications. The site webifies existing techniques (e.g. acceptance sampling) as well as new theoretical derivations on runs and scans. The website is mentioned and pointed to in many sites such as the NIST/SEMATECH Online Engineering Statistics Handbook.
I currently teach courses in
Statistics and Data Analysis for MBA students and Business undergraduates at the
Smith School of Business. I have been teaching courses
in statistics since 1994, redesigned and updated two courses, and received
awards for excellence in teaching. My teaching career kicked off as a teaching
assistant during my B.A. and M.Sc. studies. During my Ph.D. studies at the
Israel Institute of Technology, I was promoted to instructor of the advanced
“Industrial Statistics” course. At Carnegie Mellon University I taught the
courses “Engineering Statistics and Quality Control” (for engineers) and
“Sampling, Surveys, and Society” (for social sciences), and instructed a
special Six Sigma session at the Center of Automated Learning and Discovery (CALD)
summer school.
My teaching philosophy is based on my own teaching experience as well as on dialogues with leading teachers and textbook writers, and on recent papers, websites, textbooks, and conferences. My current endeavor is the integration of a new classroom technology for interactive learning ("clickers") into Smith courses. After successful experience in two of my classes in Spring 04, we will be piloting the technology in multiple Smith courses in Fall 04.
My principles of teaching
are:
For
list of courses, awards, and more see my resume.
