Regular Decomposition of Multivariate Time Series and Other Matrices

Research output: Chapter in Book/Report/Conference proceedingChapter or book articleProfessional

6 Citations (Scopus)

Abstract

We describe and illustrate a novel algorithm for clustering a large number of time series into few 'regular groups'. Our method is inspired by the famous Szemerédi's Regularity Lemma (SRL) in graph theory. SRL suggests that large graphs and matrices can be naturally 'compressed' by partitioning elements in a small number of sets. These sets and the patterns of relations between them present a kind of structure of large objects while the more detailed structure is random-like. We develop a maximum likelihood method for finding such 'regular structures' and applied it to the case of smart meter data of households. The resulting structure appears as more informative than a structure found by k-means. The algorithm scales well with data size and the structure itself becomes more apparent with bigger data size. Therefore, our method could be useful in a broader context of emerging big data
Original languageEnglish
Title of host publicationStructural, Syntactic, and Statistical Pattern Recognition
Subtitle of host publicationJoint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings
EditorsPasi Fränti, Gavin Brown, Marco Loog, Francisco Escolano, Marcello Pelillo
Place of PublicationBerlin-Heidelberg
PublisherSpringer
Pages424-433
ISBN (Print)978-3-662-44414-6
DOIs
Publication statusPublished - 2014
MoE publication typeD2 Article in professional manuals or guides or professional information systems or text book material

Publication series

SeriesLecture Notes in Computer Science
Volume8621

Fingerprint

Time series
Smart meters
Decomposition
Graph theory
Maximum likelihood
Big data

Cite this

Reittu, H., Weiss, R., & Bazso, F. (2014). Regular Decomposition of Multivariate Time Series and Other Matrices. In P. Fränti, G. Brown, M. Loog, F. Escolano, & M. Pelillo (Eds.), Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings (pp. 424-433). Berlin-Heidelberg: Springer. Lecture Notes in Computer Science, Vol.. 8621 https://doi.org/10.1007/978-3-662-44415-3
Reittu, Hannu ; Weiss, Robert ; Bazso, F. / Regular Decomposition of Multivariate Time Series and Other Matrices. Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings. editor / Pasi Fränti ; Gavin Brown ; Marco Loog ; Francisco Escolano ; Marcello Pelillo. Berlin-Heidelberg : Springer, 2014. pp. 424-433 (Lecture Notes in Computer Science, Vol. 8621).
@inbook{b25d5e9ca4124f49a4a3187852cbb107,
title = "Regular Decomposition of Multivariate Time Series and Other Matrices",
abstract = "We describe and illustrate a novel algorithm for clustering a large number of time series into few 'regular groups'. Our method is inspired by the famous Szemer{\'e}di's Regularity Lemma (SRL) in graph theory. SRL suggests that large graphs and matrices can be naturally 'compressed' by partitioning elements in a small number of sets. These sets and the patterns of relations between them present a kind of structure of large objects while the more detailed structure is random-like. We develop a maximum likelihood method for finding such 'regular structures' and applied it to the case of smart meter data of households. The resulting structure appears as more informative than a structure found by k-means. The algorithm scales well with data size and the structure itself becomes more apparent with bigger data size. Therefore, our method could be useful in a broader context of emerging big data",
author = "Hannu Reittu and Robert Weiss and F Bazso",
note = "Project code: 86270-1.4.2 InGRID_IND-COMPET: Data analysi",
year = "2014",
doi = "10.1007/978-3-662-44415-3",
language = "English",
isbn = "978-3-662-44414-6",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "424--433",
editor = "Pasi Fr{\"a}nti and Gavin Brown and Marco Loog and Francisco Escolano and { Pelillo}, Marcello",
booktitle = "Structural, Syntactic, and Statistical Pattern Recognition",
address = "Germany",

}

Reittu, H, Weiss, R & Bazso, F 2014, Regular Decomposition of Multivariate Time Series and Other Matrices. in P Fränti, G Brown, M Loog, F Escolano & M Pelillo (eds), Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings. Springer, Berlin-Heidelberg, Lecture Notes in Computer Science, vol. 8621, pp. 424-433. https://doi.org/10.1007/978-3-662-44415-3

Regular Decomposition of Multivariate Time Series and Other Matrices. / Reittu, Hannu; Weiss, Robert; Bazso, F.

Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings. ed. / Pasi Fränti; Gavin Brown; Marco Loog; Francisco Escolano; Marcello Pelillo. Berlin-Heidelberg : Springer, 2014. p. 424-433 (Lecture Notes in Computer Science, Vol. 8621).

Research output: Chapter in Book/Report/Conference proceedingChapter or book articleProfessional

TY - CHAP

T1 - Regular Decomposition of Multivariate Time Series and Other Matrices

AU - Reittu, Hannu

AU - Weiss, Robert

AU - Bazso, F

N1 - Project code: 86270-1.4.2 InGRID_IND-COMPET: Data analysi

PY - 2014

Y1 - 2014

N2 - We describe and illustrate a novel algorithm for clustering a large number of time series into few 'regular groups'. Our method is inspired by the famous Szemerédi's Regularity Lemma (SRL) in graph theory. SRL suggests that large graphs and matrices can be naturally 'compressed' by partitioning elements in a small number of sets. These sets and the patterns of relations between them present a kind of structure of large objects while the more detailed structure is random-like. We develop a maximum likelihood method for finding such 'regular structures' and applied it to the case of smart meter data of households. The resulting structure appears as more informative than a structure found by k-means. The algorithm scales well with data size and the structure itself becomes more apparent with bigger data size. Therefore, our method could be useful in a broader context of emerging big data

AB - We describe and illustrate a novel algorithm for clustering a large number of time series into few 'regular groups'. Our method is inspired by the famous Szemerédi's Regularity Lemma (SRL) in graph theory. SRL suggests that large graphs and matrices can be naturally 'compressed' by partitioning elements in a small number of sets. These sets and the patterns of relations between them present a kind of structure of large objects while the more detailed structure is random-like. We develop a maximum likelihood method for finding such 'regular structures' and applied it to the case of smart meter data of households. The resulting structure appears as more informative than a structure found by k-means. The algorithm scales well with data size and the structure itself becomes more apparent with bigger data size. Therefore, our method could be useful in a broader context of emerging big data

U2 - 10.1007/978-3-662-44415-3

DO - 10.1007/978-3-662-44415-3

M3 - Chapter or book article

SN - 978-3-662-44414-6

T3 - Lecture Notes in Computer Science

SP - 424

EP - 433

BT - Structural, Syntactic, and Statistical Pattern Recognition

A2 - Fränti, Pasi

A2 - Brown, Gavin

A2 - Loog, Marco

A2 - Escolano, Francisco

A2 - Pelillo, Marcello

PB - Springer

CY - Berlin-Heidelberg

ER -

Reittu H, Weiss R, Bazso F. Regular Decomposition of Multivariate Time Series and Other Matrices. In Fränti P, Brown G, Loog M, Escolano F, Pelillo M, editors, Structural, Syntactic, and Statistical Pattern Recognition: Joint IAPR International Workshop, S+SSPR 2014, Joensuu, Finland, August 20-22, 2014. Proceedings. Berlin-Heidelberg: Springer. 2014. p. 424-433. (Lecture Notes in Computer Science, Vol. 8621). https://doi.org/10.1007/978-3-662-44415-3