Sparse-mode Dynamic Mode Decomposition for Disambiguating Local and Global Structures

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Ichinaga, Sara M., Brunton, Steven L., Aravkin, Aleksandr Y., Kutz, J. Nathan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911078302089216
author Ichinaga, Sara M.
Brunton, Steven L.
Aravkin, Aleksandr Y.
Kutz, J. Nathan
author_facet Ichinaga, Sara M.
Brunton, Steven L.
Aravkin, Aleksandr Y.
Kutz, J. Nathan
contents The dynamic mode decomposition (DMD) is a data-driven approach that extracts the dominant features from spatiotemporal data. In this work, we introduce sparse-mode DMD, a new variant of the optimized DMD framework that specifically leverages sparsity-promoting regularization in order to approximate DMD modes which have localized spatial structure. The algorithm maintains the noise-robust properties of optimized DMD while disambiguating between modes which are spatially local versus global in nature. In many applications, such modes are associated with discrete and continuous spectra respectively, thus allowing the algorithm to explicitly construct, in an unsupervised manner, the distinct portions of the spectrum. We demonstrate this by analyzing synthetic and real-world systems, including examples from optical waveguides, quantum mechanics, and sea surface temperature data.
format Preprint
id arxiv_https___arxiv_org_abs_2507_19787
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sparse-mode Dynamic Mode Decomposition for Disambiguating Local and Global Structures
Ichinaga, Sara M.
Brunton, Steven L.
Aravkin, Aleksandr Y.
Kutz, J. Nathan
Machine Learning
The dynamic mode decomposition (DMD) is a data-driven approach that extracts the dominant features from spatiotemporal data. In this work, we introduce sparse-mode DMD, a new variant of the optimized DMD framework that specifically leverages sparsity-promoting regularization in order to approximate DMD modes which have localized spatial structure. The algorithm maintains the noise-robust properties of optimized DMD while disambiguating between modes which are spatially local versus global in nature. In many applications, such modes are associated with discrete and continuous spectra respectively, thus allowing the algorithm to explicitly construct, in an unsupervised manner, the distinct portions of the spectrum. We demonstrate this by analyzing synthetic and real-world systems, including examples from optical waveguides, quantum mechanics, and sea surface temperature data.
title Sparse-mode Dynamic Mode Decomposition for Disambiguating Local and Global Structures
topic Machine Learning
url https://arxiv.org/abs/2507.19787