Matt Taddy  

matt.taddy at chicagogsb.edu

Assistant Professor
University of Chicago
Graduate School of Business

I have moved to a faculty position in the statistics and econometrics group at the University of Chicago Graduate School of Business. This site will soon disappear, and my new webpage is at faculty.chicagogsb.edu/matt.taddy

I successfully defended my Ph.D. thesis in June at UC Santa Cruz. This thesis, written under advisors Athanasios Kottas and Herbie Lee, is concerned with the analysis of conditional distributions through the use of Dirichlet process mixture models. Applications include quantile regression, switching regression, survival analysis, and inference about point processes. I have also been working on the treed Gaussian process software developed by Bobby Gramacy, particularly in the context of optimization and sensitivity analysis. Undergraduate and masters degrees were completed at McGill University in Montréal, and my MSc research with Russ Steele involved model selection and likelihood integration for Bayesian neural networks.

Recent technical reports (updated 28/1/08).

  • Bayesian nonparametric modeling for Markov switching regression, Taddy and Kottas.
  • A nonparametric model-based approach to inference for quantile regression, Taddy and Kottas.
  • Bayesian guided pattern search for robust local optimization, Taddy, Lee, Gray, and Griffin.
  • A statistical framework for the sensitivity analysis of radiative transfer models, Morris, Kottas, Taddy, Furfaro, and Ganapol.
  • Fast Bayesian inference for computer simulation inverse problems, Taddy, Lee, and Sansó.




    • Lincoln Park, Chicago. 15/2/08

    Bormio, Italy. 11/1/08