Reservoir time series analysis: developing reservoir computing techniques for nonlinear time series analysis from foundations to application

Braden Thorne

doi:10.26182/a427-w640

Reservoir time series analysis: developing reservoir computing techniques for nonlinear time series analysis from foundations to application

Braden Thorne

Research output: Thesis › Doctoral Thesis

119 Downloads (Pure)

Abstract

This thesis bridges the gap from reservoir computing (RC) to unsupervised learning tasks.
We address what unsupervised RC entails by introducing reservoir time series analysis (RTSA), a
novel nonlinear time series analysis field which encapsulates features generated from the reservoir
state space.
We introduce and explore various RTSA features, demonstrating them on various dynamical
systems.
We then verify the use of RTSA in practice through application to unsupervised learning tasks,
specifically signal distinction, recurrence preservation and concept drift detection.
We demonstrate strong performance by the RTSA methods across these tasks which encourages
future development of the RTSA field.

Original language	English
Qualification	Doctor of Philosophy
Awarding Institution	The University of Western Australia
Supervisors/Advisors	Correa, Debora, Supervisor Small, Michael, Supervisor Zaitouny, Ayham, Supervisor
Thesis sponsors	Australian Government Research Training Program
Award date	9 Oct 2024
DOIs	https://doi.org/10.26182/a427-w640
Publication status	Unpublished - 2024

Access to Document

10.26182/a427-w640

THESIS - DOCTOR OF PHILOSOPHY - THORNE Braden John - 2024__
This work is protected by Copyright. You may print or download ONE copy of this document for the purpose of your own non-commercial research or study. Any other use requires permission from the copyright owner. The Copyright Act requires you to attribute any copyright works you quote or paraphrase.

Cite this

@phdthesis{f38feb47fe9f418eaf151d24d38196f8,

title = "Reservoir time series analysis: developing reservoir computing techniques for nonlinear time series analysis from foundations to application",

abstract = "This thesis bridges the gap from reservoir computing (RC) to unsupervised learning tasks.We address what unsupervised RC entails by introducing reservoir time series analysis (RTSA), anovel nonlinear time series analysis field which encapsulates features generated from the reservoirstate space.We introduce and explore various RTSA features, demonstrating them on various dynamicalsystems.We then verify the use of RTSA in practice through application to unsupervised learning tasks,specifically signal distinction, recurrence preservation and concept drift detection.We demonstrate strong performance by the RTSA methods across these tasks which encouragesfuture development of the RTSA field.",

keywords = "Reservoir Computing, Nonlinear Time Series Analysis, Dynamical Systems",

author = "Braden Thorne",

year = "2024",

doi = "10.26182/a427-w640",

language = "English",

school = "The University of Western Australia",

}

TY - BOOK

T1 - Reservoir time series analysis: developing reservoir computing techniques for nonlinear time series analysis from foundations to application

AU - Thorne, Braden

PY - 2024

Y1 - 2024

N2 - This thesis bridges the gap from reservoir computing (RC) to unsupervised learning tasks.We address what unsupervised RC entails by introducing reservoir time series analysis (RTSA), anovel nonlinear time series analysis field which encapsulates features generated from the reservoirstate space.We introduce and explore various RTSA features, demonstrating them on various dynamicalsystems.We then verify the use of RTSA in practice through application to unsupervised learning tasks,specifically signal distinction, recurrence preservation and concept drift detection.We demonstrate strong performance by the RTSA methods across these tasks which encouragesfuture development of the RTSA field.

AB - This thesis bridges the gap from reservoir computing (RC) to unsupervised learning tasks.We address what unsupervised RC entails by introducing reservoir time series analysis (RTSA), anovel nonlinear time series analysis field which encapsulates features generated from the reservoirstate space.We introduce and explore various RTSA features, demonstrating them on various dynamicalsystems.We then verify the use of RTSA in practice through application to unsupervised learning tasks,specifically signal distinction, recurrence preservation and concept drift detection.We demonstrate strong performance by the RTSA methods across these tasks which encouragesfuture development of the RTSA field.

KW - Reservoir Computing

KW - Nonlinear Time Series Analysis

KW - Dynamical Systems

U2 - 10.26182/a427-w640

DO - 10.26182/a427-w640

M3 - Doctoral Thesis

ER -

Reservoir time series analysis: developing reservoir computing techniques for nonlinear time series analysis from foundations to application

Abstract

Access to Document

Fingerprint

Cite this