Allen, Theodore (Ohio State University)

Theodore Allen
1971 Neil Avenue -210 Baker Systems
Columbus, OH 43210
USA

Phone: 614-668-4769
Fax: 614-292-1793
E-mail: allen.515@osu.edu
Website: http://www-iwse.eng.ohio-state.edu/ISEFaculty/allen/allen.htm

Topics:

Monitoring Using Free-Style Text
Massive databases with free-style text fields are a common feature of virtually all types of organizations from hospitals to aviation companies to governmental security agencies. Perhaps the most promising approaches for intelligent, automatic text analysis are called "topic models". The talk begins with a review of the basics of search engines and hierarchical Bayesian models. Then, we propose the so-called topic model process control (TMPC) method for control charting systems characterized by free-style text. This approach identifies new trends and assignable causes. It is based on our own recently introduced topic supervision using "anti-words'' (SAW) and chi-squared tests on the empirical topic percentages from Markov chain Monte Carlo (MCMC). We illustrate the benefits of TMPC through its application to monitoring demo download comments for an e-company. Interesting rare, unexpected topics are discovered after supervision representing new classes of customers, and the T!
MPC method correctly indicates unusual activities and their root causes. (Intermediate)

Voting Systems Design with Applications
Voting systems only make headlines when something goes wrong. For example, there might be long lines for only one type of person influencing election results as in Ohio in 2004. We describe the many threats to voting systems, our experiences consulting with them, and our work with enlightened leaders to avoid problems. Examples include applications in Ohio and New York of simulation to avoid lines and provisioning machines and FMEA to prioritize threats. (Elementary)

The Philosophy of Probability in OR and "Blying"
The Bayesian versus Frequentist versus others debate continues and it has practical consequences for OR. Through studying example applications of optimal experimental planning, markov decision processes, and forecasting insights emerge about meta-uncertainties and mistakes that can lead to undesirable solutions. Also, the subjective nature of probability can cause, particularly in politics, optimization over probabilities and the "blying" phenomenon. (Elementary)

What OR People Should Know About Six Sigma
Six sigma continues to be a powerful force in industry and the way many people are exposed to OR related concepts. Through studying hundreds of related papers, patterns have emerged about the relationship of six sigma and academic pursuits. This talk reviews the literature and six sigma concepts from and OR point of view. It points out the consistent theme that six sigma largely relates to a specific method for managing small projects with often positive results. Data is described about 39 actual projects, what techniques were used, and what were the results. (Elementary)

Background:

Phd. Industrial & Operations Engineering UMich. (Ann Arbor)
MS Physics UCLA
BA Physics Princeton

Theodore Allen is an associate professor in Industrial & Systems engineering at the Ohio State University. He has over 40 refereed publications in applied statistics and operations research including a textbook on engineering statistics. He has won three awards for his teaching and conducted consulting on behalf of many organizations including Sagata Ltd., a statistical software and consulting company that he co-manages.