English

Optimal quantile estimation: beyond the comparison model

Data Structures and Algorithms 2024-04-08 v1

Abstract

Estimating quantiles is one of the foundational problems of data sketching. Given nn elements x1,x2,,xnx_1, x_2, \dots, x_n from some universe of size UU arriving in a data stream, a quantile sketch estimates the rank of any element with additive error at most εn\varepsilon n. A low-space algorithm solving this task has applications in database systems, network measurement, load balancing, and many other practical scenarios. Current quantile estimation algorithms described as optimal include the GK sketch (Greenwald and Khanna 2001) using O(ε1logn)O(\varepsilon^{-1} \log n) words (deterministic) and the KLL sketch (Karnin, Lang, and Liberty 2016) using O(ε1loglog(1/δ))O(\varepsilon^{-1} \log\log(1/\delta)) words (randomized, with failure probability δ\delta). However, both algorithms are only optimal in the comparison-based model, whereas most typical applications involve streams of integers that the sketch can use aside from making comparisons. If we go beyond the comparison-based model, the deterministic q-digest sketch (Shrivastava, Buragohain, Agrawal, and Suri 2004) achieves a space complexity of O(ε1logU)O(\varepsilon^{-1}\log U) words, which is incomparable to the previously-mentioned sketches. It has long been asked whether there is a quantile sketch using O(ε1)O(\varepsilon^{-1}) words of space (which is optimal as long as npoly(U)n \leq \mathrm{poly}(U)). In this work, we present a deterministic algorithm using O(ε1)O(\varepsilon^{-1}) words, resolving this line of work.

Keywords

Cite

@article{arxiv.2404.03847,
  title  = {Optimal quantile estimation: beyond the comparison model},
  author = {Meghal Gupta and Mihir Singhal and Hongxun Wu},
  journal= {arXiv preprint arXiv:2404.03847},
  year   = {2024}
}
R2 v1 2026-06-28T15:44:45.261Z