Matt Taddy
Matt Taddy



My research is directed at methodology in statistics, econometrics, and machine learning, and applications in business, social science, and engineering.

GitHub and Google Scholar

Recent slides: Empirical BFs, HTE in DE, measuring rhetoric, the gamma lasso

Select papers

Document classification by inversion of distributed language representations

Bayesian and empirical Bayesian forests with Chen, Yu, and Wyle. To appear in ICML 2015.

Heterogeneous treatment effects in digital experimentation with Gardner, Chen, and Draper.

Distributed multinomial regression, to appear in the Annals of Applied Statistics (distrom R package, +textir for use in MNIR).

One-step estimator paths for concave regularization (gamlr R package).

Multinomial inverse regression for text analysis, with discussion and rejoinder. Journal of the American Statistical Association 108, 2013. (textir R package, arXiv paper and rejoinder).

Measuring political sentiment on Twitter: factor-optimal design for multinomial inverse regression. Technometrics 55, 2013. (arXiv)

Estimating Player Contribution in Hockey with Regularized Logistic Regression, with Gramacy and Jensen. Journal of Quantitative Analysis of Sports 9, 2013. (blog, arXiv, Booth piece, Chance article).

Variable Selection and Sensitivity Analysis via Dynamic Trees with an Application to Computer Code Performance Tuning, with Gramacy and Wild. Annals of Applied Statistics 7, 2013. (arXiv). Argonne write-up.

On Estimation and Selection for Topic Models. AISTATS 2012, JMLR W&CP 22. (maptpx R package, we8there.R example).

Mixture Modelling for Marked Poisson Processes, with Kottas. Bayesian Analysis 7, 2012.

Dynamic Trees for Learning and Design, with Gramacy and Polson. Journal of the American Statistical Association 106, 2011. (arXiv, dynaTree R package)

An auto-regressive mixture model for dynamic spatial Poisson processes: Application to tracking the intensity of violent crime. Journal of the American Statistical Association 105, 2010. (local copy)

Particle learning for general mixtures, with Carvalho, Lopes, and Polson. Bayesian Analysis 5, 2010.

A Bayesian nonparametric approach to inference for quantile regression, with Kottas. Journal of Business and Economic Statistics 28, 2010. (local copy)

Designing and anlayzing a circuit device experiment using treed Gaussian processes, with Lee, Gramacy, and Gray. A version of this appears as a chapter in the Handbook of Applied Bayesian Analysis, OUP 2010.

Categorical inputs, sensitivity analysis, optimization and importance tempering with tgp version 2, with Gramacy. Journal of Statistical Software 33, 2010. (R-vignette version)

Selection of a representative sample, with Lee and Gray. Journal of Classification 27, 2010. (local copy)

Markov switching Dirichlet process mixture regression, with Kottas. Bayesian Analysis 4, 2009.

Bayesian guided pattern search for robust local optimization, with Lee, Gray, and Griffin. Technometrics 51, 2009. (local copy)

Fast inference for statistical inverse problems, with Lee and Sansó. Inverse Problems 25, 2009. (local copy)

A statistical framework for the sensitivity analysis of radiative transfer models, with Morris, Kottas, Furfaro, and Ganapol. IEEE Transactions on Geoscience and Remote Sensing 12, 2008. (local copy)

Thesis: Bayesian nonparametric analysis of conditional distributions and inference for Poisson point processes.

Booth Homepage | Booth Portal | UC Homepage Copyright © 2008 University of Chicago Booth School of Business