Targeted Learning Workshops at 2019 Deming Conference on Applied Statistics

Welcome to the GitHub repository accompanying the half-day tutorial and 2-day short course on Targeted Learning, given at the Deming Conference on Applied Statistics, December 4-6, 2019.

Half-Day Tutorial on December 4, 2019

• 9:00A-10:30A: Overview of Targeted (Machine) Learning
• 10:30A-10:45P: Break
• 10:45A-12:00P: TMLE for the Causal Impact of a Single Time-Point Intervention on Survival with Software Exercise in R

2-Day Short Course on December 5-6, 2019

Day 1: December 5, 2019

• 8:00A-9:30A: Overview of Targeted (Machine) Learning
• 9:30A-9:50A: Break
• 9:50A-11:20P: Causal Inference and Interventions
• 11:20A-12:40P: Lunch
• 12:40P-2:10P: Super (Machine) Learning and Targeted Minimum Loss-Based Estimation
• 2:10P-2:30P: Break
• 2:30P-4:10P: Super Learning in the tlverse software ecosystem
• 4:10P-4:40P: Break
• 4:40P-5:50P: Super Learning in the tlverse software ecosystem

Day 2: December 6, 2019

• 8:00A-9:30A: Targeted Maximum Likelihood Estimation
• 9:30A-10:00A: Break
• 10:00A-11:20A: Targeted Maximum Likelihood Estimation of the Average Treatment Effect in the tlverse software ecosystem
• 11:20A-11:40A: Lunch
• 11:40A-1:10P: Targeted Minimum Loss-Based Estimation of Time-to-Event Outcomes with MOSS and Targeted Minimum Loss-Based Estimation for Longitudinal Data with ltmle

Most of the software materials are based on an early draft of the (eventually forthcoming) book The Hitchhiker's Guide to the tlverse, or a Targeted Learning Practitioner's Handbook, by Mark van der Laan, Alan Hubbard, Jeremy Coyle, Nima Hejazi, Ivana Malenica, and Rachael Phillips. For a work-in-progress draft of the unabridged book, please see https://tlverse.org/tlverse-handbook or the associated GitHub repository: https://github.com/tlverse/tlverse-handbook.

Note: travis-ci currently deploys the developmental version of this workshop handbook into the gh-pages branch of this repository, which is hosted by GitHub Pages at https://tlverse.org/deming2019-workshop.

The software materials are automatically built and deployed using Binder, which supports using R and RStudio, with libraries pinned to a specific snapshot on MRAN. An RStudio session, pre-loaded with all the materials, is available via