Paper detail

Taylor Polynomial Estimator for Estimating Frequency Moments

We present a randomized algorithm for estimating the $p$th moment $F_p$ of the frequency vector of a data stream in the general update (turnstile) model to within a multiplicative factor of $1 \pm ε$, for $p > 2$, with high constant confidence. For $0 < ε\le 1$, the algorithm uses space $O( n^{1-2/p} ε^{-2} + n^{1-2/p} ε^{-4/p} \log (n))$ words. This improves over the current bound of $O(n^{1-2/p} ε^{-2-4/p} \log (n))$ words by Andoni et. al. in \cite{ako:arxiv10}. Our space upper bound matches the lower bound of Li and Woodruff \cite{liwood:random13} for $ε= (\log (n))^{-Ω(1)}$ and the lower bound of Andoni et. al. \cite{anpw:icalp13} for $ε= Ω(1)$.