Skip to content

Latest commit

 

History

History
19 lines (15 loc) · 701 Bytes

README.md

File metadata and controls

19 lines (15 loc) · 701 Bytes

Interactive identification of individual treatment effect

This code is an accompaniment to the paper titled "Interactive identification of individual treatment effect with FDR control", and produce all plots in the paper.

Overview

We propose methods to identify subjects with positive effects, with a finite-sample error control on the proportion of false identification, in a randomized experiment.

To reproduce the experiments:

$Rscript code/reproduce.R

To generate the figures:

$Rscript code/plots.R

Dependencies

The code was tested using R (version 3.6.0) and the following packages: magrittr, splines, ggplot2, randomForest, caTools, dplyr, tidyr, doParallel, foreach