Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting

Abstract Singular spectrum analysis (SSA) or extended empirical orthogonal function methods are powerful, commonly used data‐driven techniques to identify modes of variability in time series and space‐time data sets. Due to the time‐lagged embedding, these methods can provide inaccurate reconstructi...

Full description

Saved in:
Bibliographic Details
Main Authors: H. Reed Ogrosky, Samuel N. Stechmann, Nan Chen, Andrew J. Majda
Format: Article
Language:English
Published: Wiley 2019-02-01
Series:Geophysical Research Letters
Online Access:https://doi.org/10.1029/2018GL081100
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850134726207078400
author H. Reed Ogrosky
Samuel N. Stechmann
Nan Chen
Andrew J. Majda
author_facet H. Reed Ogrosky
Samuel N. Stechmann
Nan Chen
Andrew J. Majda
author_sort H. Reed Ogrosky
collection DOAJ
description Abstract Singular spectrum analysis (SSA) or extended empirical orthogonal function methods are powerful, commonly used data‐driven techniques to identify modes of variability in time series and space‐time data sets. Due to the time‐lagged embedding, these methods can provide inaccurate reconstructions of leading modes near the endpoints, which can hinder the use of these methods in real time. A modified version of the traditional SSA algorithm, referred to as SSA with conditional predictions (SSA‐CP), is presented to address these issues. It is tested on low‐dimensional, approximately Gaussian data, high‐dimensional non‐Gaussian data, and partially observed data from a multiscale model. In each case, SSA‐CP provides a more accurate real‐time estimate of the leading modes of variability than the traditional reconstruction. SSA‐CP also provides predictions of the leading modes and is easy to implement. SSA‐CP is optimal in the case of Gaussian data, and the uncertainty in real‐time estimates of leading modes is easily quantified.
format Article
id doaj-art-0e1b2d6ea10a4cc18696493d8e35e75b
institution OA Journals
issn 0094-8276
1944-8007
language English
publishDate 2019-02-01
publisher Wiley
record_format Article
series Geophysical Research Letters
spelling doaj-art-0e1b2d6ea10a4cc18696493d8e35e75b2025-08-20T02:31:39ZengWileyGeophysical Research Letters0094-82761944-80072019-02-014631851186010.1029/2018GL081100Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and ForecastingH. Reed Ogrosky0Samuel N. Stechmann1Nan Chen2Andrew J. Majda3Department of Mathematics and Applied Mathematics Virginia Commonwealth University Richmond VA USADepartment of Mathematics University of Wisconsin‐Madison Madison WI USADepartment of Mathematics University of Wisconsin‐Madison Madison WI USADepartment of Mathematics and Center for Atmosphere Ocean Science, Courant Institute of Mathematical Sciences New York University New York NY USAAbstract Singular spectrum analysis (SSA) or extended empirical orthogonal function methods are powerful, commonly used data‐driven techniques to identify modes of variability in time series and space‐time data sets. Due to the time‐lagged embedding, these methods can provide inaccurate reconstructions of leading modes near the endpoints, which can hinder the use of these methods in real time. A modified version of the traditional SSA algorithm, referred to as SSA with conditional predictions (SSA‐CP), is presented to address these issues. It is tested on low‐dimensional, approximately Gaussian data, high‐dimensional non‐Gaussian data, and partially observed data from a multiscale model. In each case, SSA‐CP provides a more accurate real‐time estimate of the leading modes of variability than the traditional reconstruction. SSA‐CP also provides predictions of the leading modes and is easy to implement. SSA‐CP is optimal in the case of Gaussian data, and the uncertainty in real‐time estimates of leading modes is easily quantified.https://doi.org/10.1029/2018GL081100
spellingShingle H. Reed Ogrosky
Samuel N. Stechmann
Nan Chen
Andrew J. Majda
Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
Geophysical Research Letters
title Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
title_full Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
title_fullStr Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
title_full_unstemmed Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
title_short Singular Spectrum Analysis With Conditional Predictions for Real‐Time State Estimation and Forecasting
title_sort singular spectrum analysis with conditional predictions for real time state estimation and forecasting
url https://doi.org/10.1029/2018GL081100
work_keys_str_mv AT hreedogrosky singularspectrumanalysiswithconditionalpredictionsforrealtimestateestimationandforecasting
AT samuelnstechmann singularspectrumanalysiswithconditionalpredictionsforrealtimestateestimationandforecasting
AT nanchen singularspectrumanalysiswithconditionalpredictionsforrealtimestateestimationandforecasting
AT andrewjmajda singularspectrumanalysiswithconditionalpredictionsforrealtimestateestimationandforecasting