Featurizing Koopman mode decomposition for robust forecasting
This article introduces an advanced Koopman mode decomposition (KMD) technique-coined Featurized Koopman Mode Decomposition (FKMD)-that uses delay embedding and a learned Mahalanobis distance to enhance analysis and prediction of high-dimensional dynamical systems. The delay embedding expands the observation space to better capture underlying manifold structures, while the Mahalanobis distance adjusts observations based on the system's dynamics. This aids in featurizing KMD in cases where good features are not a priori known. We show that FKMD improves predictions for a high-dimensional linear oscillator, a high-dimensional Lorenz attractor that is partially observed, and a cell signaling problem from cancer research.
Authors
- Jeremy Copperman (ORCID: https://orcid.org/0000-0002-5202-0690)
- Nathan Mankovich (ORCID: https://orcid.org/0000-0002-5273-6349)
- Alexander E. Davies (ORCID: https://orcid.org/0000-0002-1917-8816)
- David Aristoff (ORCID: https://orcid.org/0000-0002-0347-0812)
Institutions
- Universitat de València (ES)
- Oregon Health & Science University (US)
- Colorado State University (US)
Publication Details
- Journal
- The Journal of Chemical Physics
- Published
- 2024-08-09
- DOI
- https://doi.org/10.1063/5.0220277
- Citations
- 2
- Primary Topic
- Model Reduction and Neural Networks
- Type
- article
- Field-Weighted Citation Impact
- 0.56
Funders
- National Science Foundation
- Generalitat Valenciana
- National Institutes of Health
- Knight Cancer Institute, Oregon Health and Science University
- Conselleria de Innovación, Universidades, Ciencia y Sociedad Digital, Generalitat Valenciana
- Office of Research Infrastructure Programs, National Institutes of Health
- Cancer Early Detection Advanced Research Center, Knight Cancer Institute, Oregon Health and Science University