Hamparsum Bozdogan is Toby McKenzie Professor in Business, Information Complexity and in Model Selection in the Department of Business Analytics and Statistics, at the Haslam College of Business of the University of Tennessee in Knoxville, Tennessee, which was granted to him effective as of July 1999 by the generous contributions of Mr. McKenzie and his family.

Ham Bozdogan joined the faculty of the University of Tennessee in the Fall of 1990. Prior to coming to the University of Tennessee he was on the faculty of the University of Virginia in the Department of Mathematics and was a Visiting Associate Professor and Research Fellow at the prestigious “Akaike’s Institute,” The Institute of Statistical Mathematics in Tokyo, Japan during 1988. During this year, he received the prestigious Research Assignment Leave Award from the Graduate School of Advanced Studies from the University of Virginia.

Ham is a nationally and internationally recognized renowned expert in the area of information- theoretic statistical modeling and model selection. In particular, on the celebrated Akaike’s (1971) Information Criterion (AIC), he has extended its range of applications broadly, and has identified and repaired its lack of consistency with a new criterion of his own which is now being used in many statistical software packages including JMP, EQS, SAS, and IBM SPSS, etc. Ham is the developer of a new model selection and validation criterion called ICOMP (ICOMP for ‘information complexity’). His new criterion for model selection cleverly seeks, through information theoretic ideas, to find a balance among badness of fit, lack of parsimony, and profusion of complexity in high-dimensional complex data structures by combining scalability properties in data mining. From this basic work, he has undertaken the technical and computational implementation of the criterion to many areas of applications. These include: choosing the number of component clusters in mixture-model cluster analysis, determining the number for factors in Frequentist and Bayesian factor analysis, dynamic econometric modeling of food consumption and demand in the U.S. and the Netherlands, detecting influential observations in vector autoregressive models, to mention a few. His results elucidate many current inferential problems in statistics in linear and nonlinear multivariate models and ill-posed problems. Many doctoral students at UT, in US, and around the world, are currently using his informational modeling and complexity criterion in their research and thesis work.

Ham has been trained in a modern-school of thought in Statistics which was pioneered by Professor Hirotugu Akaike in Japan in 1971, a world renowned Japanese Statistician. Because of this, he has been often labeled as a new breed of “Informational-Japanese-School” Statistician. Ham’s current research innovations during the past decade, has triggered numerous practical applications in science, engineering, business, and in healthcare analytics and medicine with significant implications in developing intelligent hybrid models between any complex modeling problem, genetic algorithms (GA’s) and his information complexity criterion as the fitness function. Coupled with this, his current research is focused in a long-standing problem of model selection under misspecification.  He is developing new techniques, which are robust and misspecification resistant. This is important because this new approach provides researchers and practitioners with knowledge of how to guard against the misspecification of the model as we actually fit and evaluate these models and guard against potential outliers in the data set. In practice, almost always researchers and practitioners alike misspecify their models for a given particular data set. In this sense these new developments and results are very important in many areas of applied and basic research (e.g., in business analytics, engineering, social and behavioral, and medical data mining, which is currently ignored. He is further developing new tools for cancer classification from gene expression data in high-dimensions for undersized samples where the covariance matrix degenerates and is not computable to reduce the dimension to accurately classify the cancerous tissues and select the best genes for treatment protocols.

Ham serves on the International Advisory Board of the Dean of the School of Business of the Istanbul University.  He is a frequent keynote speaker at national and international conferences, and he is on program committees in many international scientific conferences. His hobbies include social networking in scientific collaboration, traveling, learning other languages and cultures.