Constant-Factor Approximation Algorithms for Socially Fair $k$-Clustering
This addresses fairness in clustering for diverse groups, offering scalable solutions with theoretical guarantees, though it builds incrementally on existing methods.
The paper tackles the socially fair clustering problem with multiple groups, presenting constant-factor approximation algorithms for variants like k-median and k-means, achieving specific approximation ratios such as (5+2√6)^p with at most k+m centers and improved practical performance on benchmark datasets.
We study approximation algorithms for the socially fair $(\ell_p, k)$-clustering problem with $m$ groups, whose special cases include the socially fair $k$-median ($p=1$) and socially fair $k$-means ($p=2$) problems. We present (1) a polynomial-time $(5+2\sqrt{6})^p$-approximation with at most $k+m$ centers (2) a $(5+2\sqrt{6}+ε)^p$-approximation with $k$ centers in time $n^{2^{O(p)}\cdot m^2}$, and (3) a $(15+6\sqrt{6})^p$ approximation with $k$ centers in time $k^{m}\cdot\text{poly}(n)$. The first result is obtained via a refinement of the iterative rounding method using a sequence of linear programs. The latter two results are obtained by converting a solution with up to $k+m$ centers to one with $k$ centers using sparsification methods for (2) and via an exhaustive search for (3). We also compare the performance of our algorithms with existing bicriteria algorithms as well as exactly $k$ center approximation algorithms on benchmark datasets, and find that our algorithms also outperform existing methods in practice.