This repo contains code to perform analyses described in our paper, "Automatic Finite-Sample Robustness Metrics," for the case of ordinary least squares and instrumental variable regressions. The repo name comes from "Z-estimator approximate maximal influence".
- You can install the R library directly from github.
> library(devtools)
> devtools::install_github("https://github.com/rgiordan/zaminfluence/",
ref="master",
subdir="zaminfluence",
force=TRUE)
You can install different branches using the ref argument.
- Run the R tests to make sure everything is working.
> devtools::test("zaminfluence")
- Done, hopefully! You should now be able to run the script in
examples/simple_examples.R.
Please submit an issue or email us if you have any questions or comments!
The original version of zaminfluence depended on a Python backend. The
Python part is no longer required.
However, if you know what you're doing and are interested in defining sensitivity to custom objectives, you may want to use Python automatic differentiation tools. To do that, follow the (more complicated) Python installation instructions. Currently the interface for custom objectives is not documented, but the authors could be prompted to write such documentation if there is a need. In other words, send us an email!