Streaming Algorithms with Large Approximation Factors
Abstract
We initiate a broad study of classical problems in the streaming model with insertions and deletions in the setting where we allow the approximation factor to be much larger than . Such algorithms can use significantly less memory than the usual setting for which for an . We study large approximations for a number of problems in sketching and streaming and the following are some of our results. For the norm/quasinorm of an -dimensional vector , , we show that obtaining a -approximation requires the same amount of memory as obtaining an -approximation for any . For estimating the norm, , we show an upper bound of bits for an -approximation, and give a matching lower bound, for almost the full range of for linear sketches. For the -heavy hitters problem, we show that the known lower bound of bits for identifying -heavy hitters holds even if we are allowed to output items that are -heavy, for almost the full range of , provided the algorithm succeeds with probability . We also obtain a lower bound for linear sketches that is tight even for constant probability algorithms. For estimating the number of distinct elements, we give an -approximation algorithm using bits of space, as well as a lower bound of bits, both excluding the storage of random bits.
Cite
@article{arxiv.2207.08075,
title = {Streaming Algorithms with Large Approximation Factors},
author = {Yi Li and Honghao Lin and David P. Woodruff and Yuheng Zhang},
journal= {arXiv preprint arXiv:2207.08075},
year = {2022}
}