As there aren't ways to check whether the algorithm is producing right results. Custom datasets need to be made and tested on the algorithm to confirm whether it is working as intended or not.
To Do
Create custom data sets and test by changing different parameters (K, lambda, fairness=False) and verify the output plots.
Description
As there aren't ways to check whether the algorithm is producing right results. Custom datasets need to be made and tested on the algorithm to confirm whether it is working as intended or not.
To Do
Create custom data sets and test by changing different parameters (K, lambda, fairness=False) and verify the output plots.
Example
Sample dataset
Protected group 1: [1, 1] [2, 1] [1.5, 2] [2.5, 2]
Protected group 2: [5, 1] [6, 1] [5.5, 2] [6.5, 2]
Visually
An unfair kmeans algorithm would output following clusters:
But a balance-notion based fair clustering algorithm would output: