An iterative rounding procedure achieves a ((3^p + 1)/2 + ε)-approximation for k-clustering under p-th power distance costs, recovering the 2-approximation for k-median and improving k-means bounds to 5+ε (metric) and 4+ε (Euclidean).
An improved greedy approximation for (metric) k-means
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DS 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
$k$-Clustering via Iterative Randomized Rounding
An iterative rounding procedure achieves a ((3^p + 1)/2 + ε)-approximation for k-clustering under p-th power distance costs, recovering the 2-approximation for k-median and improving k-means bounds to 5+ε (metric) and 4+ε (Euclidean).