This repository contains code for producing the simulations and plots in my master thesis: Diagnostic Tool for Interference Features in GATE Estimation.
The code available at https://github.com/ajchin/regression-adjustments was used as a starting point for our implementations and simulations. The repository contains code used to produce the results shown in A. Chin, Regression adjustments for estimating the global treatment effect in experiments with interference, Journal of Causal Inference, May 2019 (https://www.degruyter.com/view/j/jci.ahead-of-print/jci-2018-0026/jci-2018-0026.xml)
The directory data
serves as a placeholder folder for the data.
facebook100
is a placeholder for the files from the Facebook100 dataset, which is publicly available from the Internet Archive at https://archive.org/details/oxford-2005-facebook-matrix and other public repositories.
The directory src
contains all the code used for the thesis report
setup.R
installs and loads all the packages used in the simulations
- The
functions
folder contains helpers classified as follows:
covariate_functions.R
contains functions for generating some possible interference features.response_functions.R
generates the outcomes for different models.data_generators.R
builds data frames containing realized samples of the data.existing_estimators.R
contains baseline estimators used in Chin (2019).proposed_estimators.R
constructs the adjustment estimators from Chin (2019).variance_estimators.R
computes the estimated variance for the proposed estimators.precompute_matrices.R
contains an intermediate function that estimates the matrices involved in the variance estimation.permutation_tests.R
contains our implementation for the proposed interference tests
- The
plots
directory contains the code used to generate the plots in the thesis report. - The
simulations_final.R
directory contains all the scripts and config files used in simulations (run on the ETH Euler cluster)
The results_final
directory is a placeholder for the simulation outputs, stored as .csv files.