Estimating treatment effects from single-arm trials via latent-variable modeling

This repo contains a reference implementation for

Estimating treatment effects from single-arm trials via latent-variable modeling
Manuel Haussmann, Tran Minh Son Le, Viivi Halla-aho, Samu Kurki, Jussi Leinonen, Miika Koskinen, Samuel Kaski, Harri Lähdesmäki
27th International Conference on Artificial Intelligence and Statistics
Paper

Usage

Public benchmark example: IHDP

The following steps run through the whole pipeline once from raw files to final predicitons

1. Run benchmark_scripts/get_ihdp1000.sh to download and extract the data
2. Run src/scripts/prep_data_ihdp_local.sh to preprocess it. (extcontcode/benchdata/ihdp/ihdp.py contains the python scripts used for preprocessing)
3. Run src/sripts/run_ihdp.sh to train models according to the specifications in a separate yaml file. (See benchruns/exp_benchmark.py for details)
4. Run src/scripts/eval_ihdp.sh for an evaluation routine

See benchruns/exp_benchmark.py for an example on how to train a model and how to use benchruns/eval_ihdp.py as a second step for evaluation.