R. Dennis Cook

Several years ago the Russians inadvertently dropped one of their satellites in central Canada. Naturally, the U.S. and Canadian governments thought it might be a good idea to recover the debris, and they launched a joint effort to do so. Finding the debris turned out to be rather difficult. Although they had an idea of its general location, it could still have been scattered over several hundred square miles. Much of the debris was eventually located by measuring the level of radiation while flying over probable locations. Unusually high radiation levels might indicate the presence of debris. Locating satellite debris is an example of a problem for which the solution hinges on identifying anomalous characteristics of data. In some cases we may have a pretty good idea of the type of anomaly to expect, as in the satellite problem, where higher radiation was the anomaly. In other cases we may not know what to expect, but we look for anomalies anyway because we do know that important scientific advances often come from the unexpected. For example, physicists trace the first evidence for the existence of quarks back to an anomaly that went unnoticed at the time, in the data from Millikan's famous oil-drop experiment. The area of statistics that deals with identifying anomalies in data is called diagnostics. I spend a lot of my time thinking about diagnostic problems because I find dealing with the unexpected to be exciting and a lot of fun. How do we look for strange behavior in data? How do we decide what is usual and what is truly unusual? What should we do when we find something unusual? Answering these kinds of questions often involves developing new mathematical tools. Graphs are also very important in my work. Developing new graphical tools is a particularly exciting area because recent advances in computer technology now allow me to display data in three dimensions. And being able to "see" data in four dimensions is not far off. I have enjoyed working with a number of M.S. and Ph.D. students on diagnostic problems. Most of these students made extensive use of the School's computing facilities. I also enjoy teaching, particularly regression and experimental design. My research has been supported by grants from the National Science Foundation, the National Institutes of Health, and the Monsanto Company.