2019/08/01 by Charles Masson, Jee E. Rim, Homin K. Lee · 3 voices · 68 citations
Computer Science · #Advanced Database Systems and Queries #Bounding overwatch #Data Stream Mining Techniques #Machine Learning and Data Classification #Property (philosophy) #Quantile #Rank (graph theory) #Sample (material) #Sketch #Variance (accounting) #cs.DB #cs.DS
paper · pdf · doi:10.14778/3352063.3352135
published in Proceedings of the VLDB Endowment 12(12), 2195-2205 (Association for Computing Machinery) · 11 pages, 11 figures, VLDB
openalex publication_date 2019/08/01 · arxiv created 2019/08/28 · arxiv updated 2019/08/29 · openalex created_date 2019/09/05 · openalex updated_date 2026/08/05
Summary statistics such as the mean and variance are easily maintained for large, distributed data streams, but order statistics (i.e., sample quantiles) can only be approximately summarized. There is extensive literature on maintaining quantile sketches where the emphasis has been on bounding the rank error of the sketch while using little memory. Unfortunately, rank error guarantees do not preclude arbitrarily large relative errors, and this often occurs in practice when the data is heavily skewed. Given the distributed nature of contemporary large-scale systems, another crucial property for quantile sketches is mergeablility, i.e., several combined sketches must be as accurate as a single sketch of the same data. We present the first fully-mergeable, relative-error quantile sketching algorithm with formal guarantees. The sketch is extremely fast and accurate, and is currently being used by Datadog at a wide-scale.