Home > Research > Publications & Outputs > Classification of non‐stationary time series
View graph of relations

Classification of non‐stationary time series

Research output: Contribution to journalJournal article

Published

Standard

Classification of non‐stationary time series. / Krzemieniewska, Karolina; Eckley, Idris; Fearnhead, Paul.

In: Stat, Vol. 3, No. 1, 2014, p. 144-157.

Research output: Contribution to journalJournal article

Harvard

APA

Vancouver

Author

Krzemieniewska, Karolina ; Eckley, Idris ; Fearnhead, Paul. / Classification of non‐stationary time series. In: Stat. 2014 ; Vol. 3, No. 1. pp. 144-157.

Bibtex

@article{7a5f7a535fb64b4497f4a1ed1dbeec3e,
title = "Classification of non‐stationary time series",
abstract = "In this paper we consider the problem of classifying non-stationary time series. The method that we introduce is based on the locally stationary wavelet paradigm and seeks to take account of the fact that there may be within-class variation in the signals being analysed. Specifically, we seek to identify the most stable spectral coefficients within each training group and use these to classify a new, previously unseen, time series. In both simulated examples and an aerosol spray example provided by an industrial collaborator, our approach is found to yield superior classification performance when compared against the current state of the art.",
keywords = "classification, locally stationary, time series, wavelets",
author = "Karolina Krzemieniewska and Idris Eckley and Paul Fearnhead",
year = "2014",
doi = "10.1002/sta4.51",
language = "English",
volume = "3",
pages = "144--157",
journal = "Stat",
issn = "2049-1573",
publisher = "Wiley-Blackwell Publishing Ltd",
number = "1",

}

RIS

TY - JOUR

T1 - Classification of non‐stationary time series

AU - Krzemieniewska, Karolina

AU - Eckley, Idris

AU - Fearnhead, Paul

PY - 2014

Y1 - 2014

N2 - In this paper we consider the problem of classifying non-stationary time series. The method that we introduce is based on the locally stationary wavelet paradigm and seeks to take account of the fact that there may be within-class variation in the signals being analysed. Specifically, we seek to identify the most stable spectral coefficients within each training group and use these to classify a new, previously unseen, time series. In both simulated examples and an aerosol spray example provided by an industrial collaborator, our approach is found to yield superior classification performance when compared against the current state of the art.

AB - In this paper we consider the problem of classifying non-stationary time series. The method that we introduce is based on the locally stationary wavelet paradigm and seeks to take account of the fact that there may be within-class variation in the signals being analysed. Specifically, we seek to identify the most stable spectral coefficients within each training group and use these to classify a new, previously unseen, time series. In both simulated examples and an aerosol spray example provided by an industrial collaborator, our approach is found to yield superior classification performance when compared against the current state of the art.

KW - classification

KW - locally stationary

KW - time series

KW - wavelets

U2 - 10.1002/sta4.51

DO - 10.1002/sta4.51

M3 - Journal article

VL - 3

SP - 144

EP - 157

JO - Stat

JF - Stat

SN - 2049-1573

IS - 1

ER -