Skip to content

This is the accompanying code repository for the ICML 2021 publication Oblivious Sketching for Logistic Regression by Alexander Munteanu, Simon Omlor and David P. Woodruff.

License

Notifications You must be signed in to change notification settings

chr-peters/oblivious-sketching-logreg

Repository files navigation

Oblivious Sketching for Logistic Regression

python-version

This is the accompanying code repository for the ICML 2021 publication Oblivious Sketching for Logistic Regression by Alexander Munteanu, Simon Omlor and David P. Woodruff.

How to install

  1. Clone the repository and navigate into the new directory

    git clone https://github.com/cxan96/oblivious-sketching-logreg
    cd oblivious-sketching-logreg
  2. Create and activate a new virtual environment

    python -m venv venv
    . ./venv/bin/activate
  3. Install the package locally

    pip install .
  4. To confirm that everything worked, install pytest and run the tests

    pip install pytest
    pytest

How to run the experiments

The scripts directory contains multiple python scripts that can be used to run the experiments. Just make sure, that everything is installed properly.

For example, to run the covertype experiments you can use the following command:

python scripts/run_experiments_covertype.py

How to recreate the plots

The plots can be recreated using the jupyter notebooks that can be found in the notebooks directory. Instructions on how to set up a jupyter environment can be found here.

About

This is the accompanying code repository for the ICML 2021 publication Oblivious Sketching for Logistic Regression by Alexander Munteanu, Simon Omlor and David P. Woodruff.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published