Clustering noisy temporal data

Research output: Book chapter/Published conference paperConference paper

Abstract

Clustering time series data is frequently hampered by var-ious noise components within the signal. These disturbances aect the ability of clustering to detect similarities across the various signals, which may result in poor clustering results. We propose a method, which rstsmooths out such noise using wavelet decomposition and thresholding, then reconstructs the original signal (with minimised noise) and nallyundertakes the clustering on this new signal. We experimentally evaluate the proposed method on 250 signals that are generated from ve classes of signals. Our proposed method achieves improved clustering results.
Original languageEnglish
Title of host publication15th International Conference on Advanced Data Mining and Applications (ADMA 2019)
PublisherSpringer
Pages185-194
Number of pages9
ISBN (Electronic)9783030352318
ISBN (Print)9783030352301
Publication statusPublished - Nov 2019
Event15th International Conference on Advanced Data Mining and Applications 2019: ADMA 2019 - Hi Chance (Dalian) Science & Technology Center, Dalian, China
Duration: 21 Nov 201923 Nov 2019
http://adma2019.neusoft.edu.cn/
https://link.springer.com/book/10.1007/978-3-030-35231-8 (proceedings)
http://adma2019.neusoft.edu.cn/wp-content/uploads/2019/11/ADMA-2019-Program_V5.02.pdf (program)

Publication series

NameLecture Notes in Computer Science
Volume11888

Conference

Conference15th International Conference on Advanced Data Mining and Applications 2019
CountryChina
CityDalian
Period21/11/1923/11/19
Internet address

Fingerprint

Wavelet decomposition
Time series

Cite this

Grant, P., & Islam, Z. (2019). Clustering noisy temporal data. In 15th International Conference on Advanced Data Mining and Applications (ADMA 2019) (pp. 185-194). (Lecture Notes in Computer Science; Vol. 11888). Springer.
Grant, Paul ; Islam, Zahid. / Clustering noisy temporal data. 15th International Conference on Advanced Data Mining and Applications (ADMA 2019). Springer, 2019. pp. 185-194 (Lecture Notes in Computer Science).
@inproceedings{ddcffe9223e14420801a4e6d0ab4e299,
title = "Clustering noisy temporal data",
abstract = "Clustering time series data is frequently hampered by var-ious noise components within the signal. These disturbances aect the ability of clustering to detect similarities across the various signals, which may result in poor clustering results. We propose a method, which rstsmooths out such noise using wavelet decomposition and thresholding, then reconstructs the original signal (with minimised noise) and nallyundertakes the clustering on this new signal. We experimentally evaluate the proposed method on 250 signals that are generated from ve classes of signals. Our proposed method achieves improved clustering results.",
author = "Paul Grant and Zahid Islam",
year = "2019",
month = "11",
language = "English",
isbn = "9783030352301",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "185--194",
booktitle = "15th International Conference on Advanced Data Mining and Applications (ADMA 2019)",
address = "United States",

}

Grant, P & Islam, Z 2019, Clustering noisy temporal data. in 15th International Conference on Advanced Data Mining and Applications (ADMA 2019). Lecture Notes in Computer Science, vol. 11888, Springer, pp. 185-194, 15th International Conference on Advanced Data Mining and Applications 2019, Dalian, China, 21/11/19.

Clustering noisy temporal data. / Grant, Paul; Islam, Zahid.

15th International Conference on Advanced Data Mining and Applications (ADMA 2019). Springer, 2019. p. 185-194 (Lecture Notes in Computer Science; Vol. 11888).

Research output: Book chapter/Published conference paperConference paper

TY - GEN

T1 - Clustering noisy temporal data

AU - Grant, Paul

AU - Islam, Zahid

PY - 2019/11

Y1 - 2019/11

N2 - Clustering time series data is frequently hampered by var-ious noise components within the signal. These disturbances aect the ability of clustering to detect similarities across the various signals, which may result in poor clustering results. We propose a method, which rstsmooths out such noise using wavelet decomposition and thresholding, then reconstructs the original signal (with minimised noise) and nallyundertakes the clustering on this new signal. We experimentally evaluate the proposed method on 250 signals that are generated from ve classes of signals. Our proposed method achieves improved clustering results.

AB - Clustering time series data is frequently hampered by var-ious noise components within the signal. These disturbances aect the ability of clustering to detect similarities across the various signals, which may result in poor clustering results. We propose a method, which rstsmooths out such noise using wavelet decomposition and thresholding, then reconstructs the original signal (with minimised noise) and nallyundertakes the clustering on this new signal. We experimentally evaluate the proposed method on 250 signals that are generated from ve classes of signals. Our proposed method achieves improved clustering results.

M3 - Conference paper

SN - 9783030352301

T3 - Lecture Notes in Computer Science

SP - 185

EP - 194

BT - 15th International Conference on Advanced Data Mining and Applications (ADMA 2019)

PB - Springer

ER -

Grant P, Islam Z. Clustering noisy temporal data. In 15th International Conference on Advanced Data Mining and Applications (ADMA 2019). Springer. 2019. p. 185-194. (Lecture Notes in Computer Science).