Robust $k$-means Clustering for Distributions with Two Moments
This work addresses robust clustering for data with limited moment assumptions, offering theoretical guarantees that are incremental but rigorous for applications in statistics and machine learning.
The paper tackles robust k-means clustering for distributions with only two bounded moments, providing non-asymptotic excess distortion bounds that extend Pollard's asymptotic result to general Hilbert spaces, with matching upper and lower bounds in ℝ^d that have sub-Gaussian form.
We consider the robust algorithms for the $k$-means clustering problem where a quantizer is constructed based on $N$ independent observations. Our main results are median of means based non-asymptotic excess distortion bounds that hold under the two bounded moments assumption in a general separable Hilbert space. In particular, our results extend the renowned asymptotic result of Pollard, 1981 who showed that the existence of two moments is sufficient for strong consistency of an empirically optimal quantizer in $\mathbb{R}^d$. In a special case of clustering in $\mathbb{R}^d$, under two bounded moments, we prove matching (up to constant factors) non-asymptotic upper and lower bounds on the excess distortion, which depend on the probability mass of the lightest cluster of an optimal quantizer. Our bounds have the sub-Gaussian form, and the proofs are based on the versions of uniform bounds for robust mean estimators.