ICML2019 Proportionally Fair Clustering

Authors: Xingyu Chen, Brandon Fain, Charles Lyu, Kamesh Munagala

This repository serves as the code base for ICML2019 Submission Paper: Proportionally Fair Clustering.

Requirements:

Python >= 3.7.0

scikit-learn >= 0.20.0

matplotlib >= 3.0.2

numpy >= 1.15.4

How to repeat our experiments

Go to root folder, run python experiment.py to start experiment. In command line, you can specify the -proportionality measure used for local capture algorithm. You can also specify which database you run your experiment with --file_name option. (available names kdd , iris , diabetes ). To repeat the three experiments we did, you can run the following three commands,

python experiment.py --file_name iris --rho 1.00001

python experiment.py --file_name diabetes --rho 1.00001

python experiment.py --file_name kdd --rho 2 --sample

To customize your own experiments, see More Help section for support.

More Help

usage: experiment.py [-h] [--sample] [--sample_clients SAMPLE_NUM]

​ [--sample_centers CENTER_NUM] [--file_name FILE_NAME]

​ [--rho RHO]

optional arguments:

-h, --help show this help message and exit

--sample

--sample_clients SAMPLE_NUM

--sample_centers CENTER_NUM

--file_name FILE_NAME

--rho RHO