Skip to content

What the Package Does (One Line, Title Case)

License

Unknown, MIT licenses found

Licenses found

Unknown
LICENSE
MIT
LICENSE.md
Notifications You must be signed in to change notification settings

shamindras/ascifit

Repository files navigation

ascifit

The goal of ascifit is to have a set of common utilities to reproduce results from the paper Adversarial Sign Corrupted Isotonic Regression.

This repository is developed and maintained by

Citation

To reference this work, please cite the latest version of the paper

@misc{shrotriya2022ascireg,
  title   = {Adversarial Sign-Corrupted Isotonic Regression},
  author  = {Shamindra Shrotriya and Matey Neykov},
  year    = 2022,
  eprint  = {arXiv:2207.07075}
}

Installation

In order to reproduce the figures you need to first install the renv package from CRAN in a separate R session.

install.packages("renv")

You can then clone the repository locally as usual, for example using SSH

git@github.com:shamindras/ascifit.git #SSH approach

Reproducibility

  1. After locally cloning the repository, you can open the ascifit.Rproj project file in RStudio. This will create a separate RStudio session.

  2. Once ascifit.Rproj is opened in RStudio, you can open the file

    experiments/R/paper_results/create-ascifit-plots.R
  3. Select all the contents, and run the entire file. It will produce both png/TeX versions of the three figures, as follows:

    out_ascifit1_50_reps.png # -> Figure 1 in paper
    sim_plt_50_reps.png # -> Figure 2 in paper

    Note: In the paper we use the .tex versions for the figures not the .png to reduce the resulting paper file size. However both the .tex and corresponding named .png files produce equivalent plots. The .png formats are typically easier to view.

About

What the Package Does (One Line, Title Case)

Resources

License

Unknown, MIT licenses found

Licenses found

Unknown
LICENSE
MIT
LICENSE.md

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published