SIRCS: Slope-intercept-residual Compression by Correlation Sequencing for Multi-stream High Variation Data

Zixin Ye, Wen Hua, Liwei Wang, Xiaofang Zhou

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

Abstract

Multi-stream data with high variation is ubiquitous in the modern network systems. With the development of telecommunication technologies, robust data compression techniques are urged to be developed. In this paper, we humbly introduce a novel technique specifically for high variation signal data: SIRCS, which applies linear regression model for slope, intercept and residual decomposition of the multi data stream and combines the advanced tree mapping techniques. SIRCS inherits the advantages from the existing grouping compression algorithms, like GAMPS. With the newly invented correlation sorting techniques: the correlation tree mapping, SIRCS can practically improve the compression ratio by 13% from the traditional clustering mapping scheme. The application of the linear model decomposition can further facilitate the improvement of the algorithm performance from the state-of-art algorithms, with the RMSE decrease 4% and the compression time dramatically drop compared to the GAMPS. With the wide range of the error tolerance from 1% to 27%, SIRCS performs consistently better than all evaluated state-of-art algorithms regarding compression efficiency and accuracy.

Original languageEnglish
Title of host publicationDatabase Systems for Advanced Applications - 24th International Conference, DASFAA 2019, Proceedings
EditorsJuggapong Natwichai, Guoliang Li, Jun Yang, Joao Gama, Yongxin Tong
PublisherSpringer Verlag
Pages191-206
Number of pages16
ISBN (Print)9783030185756
DOIs
Publication statusPublished - 2019
Externally publishedYes
Event24th International Conference on Database Systems for Advanced Applications, DASFAA 2019 - Chiang Mai, Thailand
Duration: 22 Apr 201925 Apr 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11446 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th International Conference on Database Systems for Advanced Applications, DASFAA 2019
Country/TerritoryThailand
CityChiang Mai
Period22/04/1925/04/19

Keywords

  • Correlation mapping
  • Error detection
  • High variation data
  • Linear regression model
  • Multi-signal compression

ASJC Scopus subject areas

  • Information Systems

Fingerprint

Dive into the research topics of 'SIRCS: Slope-intercept-residual Compression by Correlation Sequencing for Multi-stream High Variation Data'. Together they form a unique fingerprint.

Cite this