Associate Professor
Website: http://www.stat.missouri.edu/~thombsl/
Areas of Research: Time Series, Resampling Methods for Correlated Variables, Statistics Education
Brief Description of Major Research Activities: My main research interest relates to the analysis of dependent data. Within this very broad area, my focus includes (i) Statistical inference for temporally-correlated data, such as linear and nonlinear time series, and (ii) Multivariate methods for classification and data mining, and (iii) general modeling of complex, time-indexed data sets. My early work (1990) in the area of bootstrapping autoregressive time series represented one of the initial efforts of bootstrapping dependent data. Recent inquiries (1996, 2001) deal with economic time series models characterized by high volatility. Such models are extremely useful in the financial area, but such behavior arises in applications throughout the physical sciences. My collaborative efforts in the multivariate area are an example of my interest in the interplay between the practical and theoretical aspects of statistics. Papers (1994) with collaborators in the field of exercise science represent a blending of the spectral analysis approach to real-time force plate sway data. My investigations of the role that multivariate methods play in business and economic applications have led to new research which investigates the performance of a generalization of the standard CART classification approach in the area of data mining. My work in statistical education (1996) with Drs. Spurrier and Edwards advocates the use of hands-on learning activities for use in elementary statistics.
Why I Like Being a Statistician: I love the interdisciplinary flavor of our field. Statistics is used in every field of inquiry, and my collaborative opportunities with other scientists have been very rewarding experiences. I also feel that that my ability to "think statistically" helps me to enjoy everyday activities such as reading the newspaper, hearing about new medical advances, or debating social issues with my friends.
Six Selected Publications:
Li, X, Sweigart, J, Teng, J,. Donohue, J. and Thombs, L.,(2001) "A Dynamic Programming Based Pruning Method for Decision Trees," INFORMS Journal on Computing, 13, 332-344.
Li, X, Sweigart, J, Teng, J,. Donohue, J. and Thombs, L., (2002). "Linear Discriminant Function Based Multivariate Decision Trees," to appear in IEEE Transactions on Systems, Man and Cybernetics, Part B.
McClenaghan, B.A.; Williams, H.; Dickerson, J. and Thombs L. A. (1994). "Use of the Spectral Signature of Postural Forces to Discriminate Perturbations in Standing Stability," Clinical Biomechanics. (9) 21-27.
Romano, J. and Thombs, L. A. (1996). "Inference for Autocorrelations Under Weak Assumptions," Journal of the American Statistical Association, 91, 590-600. .
Spurrier, J.D., Edwards, D.G., and Thombs, L.A. (1999). Statistics: Learning by Doing, Whittier Press.
Thombs, L. A. and Schucany, W. R. (1990). "Bootstrap Prediction Intervals for Autoregression," Journal of the American Statistical Association, 85, 486 - 492.