DETECTING DETERMINISM IN TIME SERIES DATA: WHEN SHOULD WE BOTHER TO BUILD MODELS?

Michael Small; Chi K. Tse

DETECTING DETERMINISM IN TIME SERIES DATA: WHEN SHOULD WE BOTHER TO BUILD MODELS?

Michael Small, Chi K. Tse

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

Abstract

Nonlinear modeling routines are often applied in an effort to extract underlying determinism from time series data. The best of these methods perform well for short noisy time series when there is determinism in the underlying system. We show that nonlinear modeling does not distinguish between a static nonlinear transformation of linearly filtered noise and dynamic nonlinearity. To relieve this problem we recommend that surrogate data methods should be applied prior to nonlinear modeling, and the results of that analysis used to guide model selection.

Original language	English
Title of host publication	Proceedings of International Symposium on Nonlinear Theory and Its Applications (NOLTA'2002)
Pages	21-24
Publication status	Published - 2002
Externally published	Yes
Event	2002 International Symposium on Nonlinear Theory and its Applications (NOLTA’2002) - Xi'an International Conference Centre, Xi'an, China Duration: 7 Oct 2002 → 11 Oct 2002 https://www.ieice.org/nolta/symposium/archive/2002/01.html

Conference

Conference	2002 International Symposium on Nonlinear Theory and its Applications (NOLTA’2002)
Place	China
City	Xi'an
Period	7/10/02 → 11/10/02
Internet address	https://www.ieice.org/nolta/symposium/archive/2002/01.html

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{acb0f4a582f643d0b1c198f2178dc0b9,

title = "DETECTING DETERMINISM IN TIME SERIES DATA: WHEN SHOULD WE BOTHER TO BUILD MODELS?",

abstract = "Nonlinear modeling routines are often applied in an effort to extract underlying determinism from time series data. The best of these methods perform well for short noisy time series when there is determinism in the underlying system. We show that nonlinear modeling does not distinguish between a static nonlinear transformation of linearly filtered noise and dynamic nonlinearity. To relieve this problem we recommend that surrogate data methods should be applied prior to nonlinear modeling, and the results of that analysis used to guide model selection.",

author = "Michael Small and Tse, {Chi K.}",

year = "2002",

language = "English",

pages = "21--24",

booktitle = "Proceedings of International Symposium on Nonlinear Theory and Its Applications (NOLTA'2002)",

note = "2002 International Symposium on Nonlinear Theory and its Applications (NOLTA{\textquoteright}2002) ; Conference date: 07-10-2002 Through 11-10-2002",

url = "https://www.ieice.org/nolta/symposium/archive/2002/01.html",

}

DETECTING DETERMINISM IN TIME SERIES DATA: WHEN SHOULD WE BOTHER TO BUILD MODELS? / Small, Michael; Tse, Chi K.
Proceedings of International Symposium on Nonlinear Theory and Its Applications (NOLTA'2002). 2002. p. 21-24.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - DETECTING DETERMINISM IN TIME SERIES DATA

T2 - 2002 International Symposium on Nonlinear Theory and its Applications (NOLTA’2002)

AU - Small, Michael

AU - Tse, Chi K.

PY - 2002

Y1 - 2002

N2 - Nonlinear modeling routines are often applied in an effort to extract underlying determinism from time series data. The best of these methods perform well for short noisy time series when there is determinism in the underlying system. We show that nonlinear modeling does not distinguish between a static nonlinear transformation of linearly filtered noise and dynamic nonlinearity. To relieve this problem we recommend that surrogate data methods should be applied prior to nonlinear modeling, and the results of that analysis used to guide model selection.

AB - Nonlinear modeling routines are often applied in an effort to extract underlying determinism from time series data. The best of these methods perform well for short noisy time series when there is determinism in the underlying system. We show that nonlinear modeling does not distinguish between a static nonlinear transformation of linearly filtered noise and dynamic nonlinearity. To relieve this problem we recommend that surrogate data methods should be applied prior to nonlinear modeling, and the results of that analysis used to guide model selection.

M3 - RGC 32 - Refereed conference paper (with host publication)

SP - 21

EP - 24

BT - Proceedings of International Symposium on Nonlinear Theory and Its Applications (NOLTA'2002)

Y2 - 7 October 2002 through 11 October 2002

ER -

DETECTING DETERMINISM IN TIME SERIES DATA: WHEN SHOULD WE BOTHER TO BUILD MODELS?

Abstract

Conference

Other Access Options

Permanent Link

Fingerprint

Cite this