Using convolution to mine obscure periodic patterns in one pass

Mohamed G. Elfeky, Walid G. Aref, Ahmed Elmagarmid

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

The mining of periodic patterns in time series databases is an interesting data mining problem that can be envisioned as a tool for forecasting and predicting the future behavior of time series data. Existing periodic patterns mining algorithms either assume that the periodic rate (or simply the period) is user-specified, or try to detect potential values for the period in a separate phase. The former assumption is a considerable disadvantage, especially in time series databases where the period is not known a priori. The latter approach results in a multi-pass algorithm, which on the other hand is to be avoided in online environments (e.g., data streams). In this paper, we develop an algorithm that mines periodic patterns in time series databases with unknown or obscure periods such that discovering the period is part of the mining process. Based on convolution, our algorithm requires only one pass over a time series of length n, with O(nlogn) time complexity.

Original languageEnglish
Pages (from-to)606-620
Number of pages15
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2992
Publication statusPublished - 1 Dec 2004
Externally publishedYes

Fingerprint

Convolution
Time series
Databases
Mining
Process Mining
Time Series Data
Data Streams
Data Mining
Time Complexity
Data mining
Forecasting
Unknown

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

@article{0c400d41c7f94c8e874d25fc6d602658,
title = "Using convolution to mine obscure periodic patterns in one pass",
abstract = "The mining of periodic patterns in time series databases is an interesting data mining problem that can be envisioned as a tool for forecasting and predicting the future behavior of time series data. Existing periodic patterns mining algorithms either assume that the periodic rate (or simply the period) is user-specified, or try to detect potential values for the period in a separate phase. The former assumption is a considerable disadvantage, especially in time series databases where the period is not known a priori. The latter approach results in a multi-pass algorithm, which on the other hand is to be avoided in online environments (e.g., data streams). In this paper, we develop an algorithm that mines periodic patterns in time series databases with unknown or obscure periods such that discovering the period is part of the mining process. Based on convolution, our algorithm requires only one pass over a time series of length n, with O(nlogn) time complexity.",
author = "Elfeky, {Mohamed G.} and Aref, {Walid G.} and Ahmed Elmagarmid",
year = "2004",
month = "12",
day = "1",
language = "English",
volume = "2992",
pages = "606--620",
journal = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Using convolution to mine obscure periodic patterns in one pass

AU - Elfeky, Mohamed G.

AU - Aref, Walid G.

AU - Elmagarmid, Ahmed

PY - 2004/12/1

Y1 - 2004/12/1

N2 - The mining of periodic patterns in time series databases is an interesting data mining problem that can be envisioned as a tool for forecasting and predicting the future behavior of time series data. Existing periodic patterns mining algorithms either assume that the periodic rate (or simply the period) is user-specified, or try to detect potential values for the period in a separate phase. The former assumption is a considerable disadvantage, especially in time series databases where the period is not known a priori. The latter approach results in a multi-pass algorithm, which on the other hand is to be avoided in online environments (e.g., data streams). In this paper, we develop an algorithm that mines periodic patterns in time series databases with unknown or obscure periods such that discovering the period is part of the mining process. Based on convolution, our algorithm requires only one pass over a time series of length n, with O(nlogn) time complexity.

AB - The mining of periodic patterns in time series databases is an interesting data mining problem that can be envisioned as a tool for forecasting and predicting the future behavior of time series data. Existing periodic patterns mining algorithms either assume that the periodic rate (or simply the period) is user-specified, or try to detect potential values for the period in a separate phase. The former assumption is a considerable disadvantage, especially in time series databases where the period is not known a priori. The latter approach results in a multi-pass algorithm, which on the other hand is to be avoided in online environments (e.g., data streams). In this paper, we develop an algorithm that mines periodic patterns in time series databases with unknown or obscure periods such that discovering the period is part of the mining process. Based on convolution, our algorithm requires only one pass over a time series of length n, with O(nlogn) time complexity.

UR - http://www.scopus.com/inward/record.url?scp=35048882067&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048882067&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:35048882067

VL - 2992

SP - 606

EP - 620

JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SN - 0302-9743

ER -