Skip to main navigation Skip to search Skip to main content

Multi-Hierarchies: Accurately computing realtime statistical measures on data streams

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Computing statistical measures is a fundamental problem for mining data streams. Sometimes user wants to query the realtime correlation of data streams. In this paper, we introduce a system for computing realtime statistical measures of data streams. The system updates the realtime summaries which are used to compute affine relationships. We process every elements in every data stream only once, and get a similar accuracy rating compared with the static methods. To the best of our knowledge, we present a new method of computing affine relationship. Our system employs the multi-Hierarchies approach in the Sliding Window Model. First, we change AFCLST Clustering algorithm. Second, the Bottom-Up Updating algorithm updates the summaries which every hierarchy has stored after the Cumulative Calculation algorithms. Third, the Query Response algorithm uses summaries to compute the statistical measure. Finally, we establish the accuracy rating of our approach by performing several experiments on real datasets.

Original languageEnglish
Title of host publicationWireless Algorithms, Systems and Applications - 9th International Conference, WASA 2014, Proceedings
EditorsZhipeng Cai, Chaokun Wang, Siyao Cheng, Hongzhi Wang, Hong Gao
PublisherSpringer Verlag
Pages729-739
Number of pages11
ISBN (Electronic)9783319077819
DOIs
StatePublished - 2014
Event9th International Conference on Wireless Algorithms, Systems and Applications, WASA 2014 - Harbin, China
Duration: 23 Jun 201425 Jun 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8491
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference9th International Conference on Wireless Algorithms, Systems and Applications, WASA 2014
Country/TerritoryChina
CityHarbin
Period23/06/1425/06/14

Fingerprint

Dive into the research topics of 'Multi-Hierarchies: Accurately computing realtime statistical measures on data streams'. Together they form a unique fingerprint.

Cite this