Overview
David Madigan is a statistician and data scientist (formerly Columbia University; later provost/academic leadership roles) known for foundational work on Bayesian model averaging, Bayesian graphical models, and large-scale observational health data analysis.
Key Contributions / Features
- Occam's window and MC³ (Madigan-Raftery 1994; Madigan-York 1995): the two core strategies for making Bayesian model averaging computationally feasible — pruning to a data-supported model set, and MCMC over model space.
- BMA tutorial (Hoeting-Madigan-Raftery-Volinsky 1999): co-author of the canonical practical reference.
- BMA for linear regression (Raftery-Madigan-Hoeting 1997): co-developer of Occam's window and MC³ for regression model averaging (Bayesian Model Averaging).
- Bayesian graphical/network models and their use in expert systems and observational studies.
Related