Multiple and partial periodicity mining in time series databases

Christos Berberidis; Walid G. Aref; Mikhail Atallah; Ioannis Vlahavas; Ahmed Khalifa Elmagarmid

Multiple and partial periodicity mining in time series databases

Christos Berberidis, Walid G. Aref, Mikhail Atallah, Ioannis Vlahavas, Ahmed Khalifa Elmagarmid

Research output: Contribution to conference › Paper › peer-review

Abstract

Periodicity search in time series is a problem that has been investigated by mathematicians in various areas, such as sta- tistics, economics, and digital signal processing. For large data- bases of time series data, scalability becomes an issue that tradi- tional techniques fail to address. In existing time series mining algorithms for detecting periodic patterns, the period length is user- specified. This is a drawback especially for datasets where no pe- riod length is known in advance. We propose an algorithm that extracts a set of candidate periods featured in a time series that satisfy a minimum confidence threshold, by utilizing the autocor- relation function and FFT as a filter. We provide some mathemati- cal background as well as experimental results.

Original language	English
Number of pages	5
Publication status	Published - 2002
Externally published	Yes
Event	ECAI 2002 - Lyon, France Duration: 21 Jul 2002 → 26 Jul 2002

Conference

Conference	ECAI 2002
Country/Territory	France
City	Lyon
Period	21/07/02 → 26/07/02

Access to Document

https://scholar.google.com/citations?view_op=view_citation&hl=en&user=Cb9BDJkAAAAJ&pagesize=100&citation_for_view=Cb9BDJkAAAAJ:TQgYirikUcIC

Cite this

@conference{caa0c16a5b7a4dd2887b8d245526b75b,

title = "Multiple and partial periodicity mining in time series databases",

abstract = "Periodicity search in time series is a problem that has been investigated by mathematicians in various areas, such as sta- tistics, economics, and digital signal processing. For large data- bases of time series data, scalability becomes an issue that tradi- tional techniques fail to address. In existing time series mining algorithms for detecting periodic patterns, the period length is user- specified. This is a drawback especially for datasets where no pe- riod length is known in advance. We propose an algorithm that extracts a set of candidate periods featured in a time series that satisfy a minimum confidence threshold, by utilizing the autocor- relation function and FFT as a filter. We provide some mathemati- cal background as well as experimental results.",

author = "Christos Berberidis and Aref, {Walid G.} and Mikhail Atallah and Ioannis Vlahavas and Elmagarmid, {Ahmed Khalifa}",

year = "2002",

language = "English",

note = "ECAI 2002 ; Conference date: 21-07-2002 Through 26-07-2002",

}

TY - CONF

T1 - Multiple and partial periodicity mining in time series databases

AU - Berberidis, Christos

AU - Aref, Walid G.

AU - Atallah, Mikhail

AU - Vlahavas, Ioannis

AU - Elmagarmid, Ahmed Khalifa

PY - 2002

Y1 - 2002

N2 - Periodicity search in time series is a problem that has been investigated by mathematicians in various areas, such as sta- tistics, economics, and digital signal processing. For large data- bases of time series data, scalability becomes an issue that tradi- tional techniques fail to address. In existing time series mining algorithms for detecting periodic patterns, the period length is user- specified. This is a drawback especially for datasets where no pe- riod length is known in advance. We propose an algorithm that extracts a set of candidate periods featured in a time series that satisfy a minimum confidence threshold, by utilizing the autocor- relation function and FFT as a filter. We provide some mathemati- cal background as well as experimental results.

AB - Periodicity search in time series is a problem that has been investigated by mathematicians in various areas, such as sta- tistics, economics, and digital signal processing. For large data- bases of time series data, scalability becomes an issue that tradi- tional techniques fail to address. In existing time series mining algorithms for detecting periodic patterns, the period length is user- specified. This is a drawback especially for datasets where no pe- riod length is known in advance. We propose an algorithm that extracts a set of candidate periods featured in a time series that satisfy a minimum confidence threshold, by utilizing the autocor- relation function and FFT as a filter. We provide some mathemati- cal background as well as experimental results.

M3 - Paper

T2 - ECAI 2002

Y2 - 21 July 2002 through 26 July 2002

ER -

Multiple and partial periodicity mining in time series databases

Abstract

Conference

Access to Document

Fingerprint

Cite this