A novel symbolic aggregate approximation for time series

Yufeng Yu, Yuelong Zhu, Dingsheng Wan, Huan Liu, Qun Zhao

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Symbolic Aggregate approximation (SAX) is a classical symbolic approach in many time series data mining applications. However, SAX only reflects the segment mean value feature and misses important information in a segment, namely the trend of the value change in the segment. Such a miss may cause a wrong classification in some cases, since the SAX representation cannot distinguish different time series with similar average values but different trends. In this paper, we present Trend Feature Symbolic Aggregate approximation (TFSAX) to solve this problem. First, we utilize Piecewise Aggregate Approximation (PAA) approach to reduce dimensionality and discretize the mean value of each segment by SAX. Second, extract trend feature in each segment by using trend distance factor and trend shape factor. Then, design multi-resolution symbolic mapping rules to discretize trend information into symbols. We also propose a modified distance measure by integrating the SAX distance with a weighted trend distance. We show that our distance measure has a tighter lower bound to the Euclidean distance than that of the original SAX. The experimental results on diverse time series data sets demonstrate that our proposed representation significantly outperforms the original SAX representation and an improved SAX representation for classification.

Original languageEnglish (US)
Title of host publicationProceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019
EditorsSukhan Lee, Hyunseung Choo, Roslan Ismail
PublisherSpringer Verlag
Pages805-822
Number of pages18
ISBN (Print)9783030190620
DOIs
StatePublished - Jan 1 2019
Event13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019 - Phuket, Thailand
Duration: Jan 4 2019Jan 6 2019

Publication series

NameAdvances in Intelligent Systems and Computing
Volume935
ISSN (Print)2194-5357

Conference

Conference13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019
CountryThailand
CityPhuket
Period1/4/191/6/19

Fingerprint

Time series
Data mining

Keywords

  • Distance measure
  • Lower bound
  • Symbolic aggregate approximation
  • Time series
  • Trend feature

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Yu, Y., Zhu, Y., Wan, D., Liu, H., & Zhao, Q. (2019). A novel symbolic aggregate approximation for time series. In S. Lee, H. Choo, & R. Ismail (Eds.), Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019 (pp. 805-822). (Advances in Intelligent Systems and Computing; Vol. 935). Springer Verlag. https://doi.org/10.1007/978-3-030-19063-7_65

A novel symbolic aggregate approximation for time series. / Yu, Yufeng; Zhu, Yuelong; Wan, Dingsheng; Liu, Huan; Zhao, Qun.

Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019. ed. / Sukhan Lee; Hyunseung Choo; Roslan Ismail. Springer Verlag, 2019. p. 805-822 (Advances in Intelligent Systems and Computing; Vol. 935).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yu, Y, Zhu, Y, Wan, D, Liu, H & Zhao, Q 2019, A novel symbolic aggregate approximation for time series. in S Lee, H Choo & R Ismail (eds), Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019. Advances in Intelligent Systems and Computing, vol. 935, Springer Verlag, pp. 805-822, 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019, Phuket, Thailand, 1/4/19. https://doi.org/10.1007/978-3-030-19063-7_65
Yu Y, Zhu Y, Wan D, Liu H, Zhao Q. A novel symbolic aggregate approximation for time series. In Lee S, Choo H, Ismail R, editors, Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019. Springer Verlag. 2019. p. 805-822. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-3-030-19063-7_65
Yu, Yufeng ; Zhu, Yuelong ; Wan, Dingsheng ; Liu, Huan ; Zhao, Qun. / A novel symbolic aggregate approximation for time series. Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019. editor / Sukhan Lee ; Hyunseung Choo ; Roslan Ismail. Springer Verlag, 2019. pp. 805-822 (Advances in Intelligent Systems and Computing).
@inproceedings{aa0a8f58b9e94e9088e00712f6078c46,
title = "A novel symbolic aggregate approximation for time series",
abstract = "Symbolic Aggregate approximation (SAX) is a classical symbolic approach in many time series data mining applications. However, SAX only reflects the segment mean value feature and misses important information in a segment, namely the trend of the value change in the segment. Such a miss may cause a wrong classification in some cases, since the SAX representation cannot distinguish different time series with similar average values but different trends. In this paper, we present Trend Feature Symbolic Aggregate approximation (TFSAX) to solve this problem. First, we utilize Piecewise Aggregate Approximation (PAA) approach to reduce dimensionality and discretize the mean value of each segment by SAX. Second, extract trend feature in each segment by using trend distance factor and trend shape factor. Then, design multi-resolution symbolic mapping rules to discretize trend information into symbols. We also propose a modified distance measure by integrating the SAX distance with a weighted trend distance. We show that our distance measure has a tighter lower bound to the Euclidean distance than that of the original SAX. The experimental results on diverse time series data sets demonstrate that our proposed representation significantly outperforms the original SAX representation and an improved SAX representation for classification.",
keywords = "Distance measure, Lower bound, Symbolic aggregate approximation, Time series, Trend feature",
author = "Yufeng Yu and Yuelong Zhu and Dingsheng Wan and Huan Liu and Qun Zhao",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-19063-7_65",
language = "English (US)",
isbn = "9783030190620",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer Verlag",
pages = "805--822",
editor = "Sukhan Lee and Hyunseung Choo and Roslan Ismail",
booktitle = "Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019",

}

TY - GEN

T1 - A novel symbolic aggregate approximation for time series

AU - Yu, Yufeng

AU - Zhu, Yuelong

AU - Wan, Dingsheng

AU - Liu, Huan

AU - Zhao, Qun

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Symbolic Aggregate approximation (SAX) is a classical symbolic approach in many time series data mining applications. However, SAX only reflects the segment mean value feature and misses important information in a segment, namely the trend of the value change in the segment. Such a miss may cause a wrong classification in some cases, since the SAX representation cannot distinguish different time series with similar average values but different trends. In this paper, we present Trend Feature Symbolic Aggregate approximation (TFSAX) to solve this problem. First, we utilize Piecewise Aggregate Approximation (PAA) approach to reduce dimensionality and discretize the mean value of each segment by SAX. Second, extract trend feature in each segment by using trend distance factor and trend shape factor. Then, design multi-resolution symbolic mapping rules to discretize trend information into symbols. We also propose a modified distance measure by integrating the SAX distance with a weighted trend distance. We show that our distance measure has a tighter lower bound to the Euclidean distance than that of the original SAX. The experimental results on diverse time series data sets demonstrate that our proposed representation significantly outperforms the original SAX representation and an improved SAX representation for classification.

AB - Symbolic Aggregate approximation (SAX) is a classical symbolic approach in many time series data mining applications. However, SAX only reflects the segment mean value feature and misses important information in a segment, namely the trend of the value change in the segment. Such a miss may cause a wrong classification in some cases, since the SAX representation cannot distinguish different time series with similar average values but different trends. In this paper, we present Trend Feature Symbolic Aggregate approximation (TFSAX) to solve this problem. First, we utilize Piecewise Aggregate Approximation (PAA) approach to reduce dimensionality and discretize the mean value of each segment by SAX. Second, extract trend feature in each segment by using trend distance factor and trend shape factor. Then, design multi-resolution symbolic mapping rules to discretize trend information into symbols. We also propose a modified distance measure by integrating the SAX distance with a weighted trend distance. We show that our distance measure has a tighter lower bound to the Euclidean distance than that of the original SAX. The experimental results on diverse time series data sets demonstrate that our proposed representation significantly outperforms the original SAX representation and an improved SAX representation for classification.

KW - Distance measure

KW - Lower bound

KW - Symbolic aggregate approximation

KW - Time series

KW - Trend feature

UR - http://www.scopus.com/inward/record.url?scp=85066821443&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066821443&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-19063-7_65

DO - 10.1007/978-3-030-19063-7_65

M3 - Conference contribution

AN - SCOPUS:85066821443

SN - 9783030190620

T3 - Advances in Intelligent Systems and Computing

SP - 805

EP - 822

BT - Proceedings of the 13th International Conference on Ubiquitous Information Management and Communication, IMCOM 2019

A2 - Lee, Sukhan

A2 - Choo, Hyunseung

A2 - Ismail, Roslan

PB - Springer Verlag

ER -