Reduced-space Gaussian Process Regression for data-driven probabilistic forecast of chaotic dynamical systems

  title={Reduced-space Gaussian Process Regression for data-driven probabilistic forecast of chaotic dynamical systems},
  author={Zhong Yi Wan and Themistoklis P. Sapsis},
  journal={Physica D: Nonlinear Phenomena},
  • Z. Y. WanT. Sapsis
  • Published 5 November 2016
  • Computer Science
  • Physica D: Nonlinear Phenomena

Time Series Forecasting Using Manifold Learning

A three-tier numerical framework based on manifold learning for the forecasting of highdimensional time series is addressed and a comparison with the Principal Component Analysis algorithm as well as with the naive random walk model and the MVAR and GPR models trained and implemented directly in the high-dimensional space are provided.

Time-series forecasting using manifold learning, radial basis function interpolation, and geometric harmonics.

We address a three-tier numerical framework based on nonlinear manifold learning for the forecasting of high-dimensional time series, relaxing the "curse of dimensionality" related to the training

Data-assisted reduced-order modeling of extreme events in complex dynamical systems

A novel hybrid framework that complements an imperfect reduced order model, with data-streams that are integrated though a recurrent neural network (RNN) architecture is developed, showing that the blended approach has improved performance compared with methods that use either data streams or the imperfect model alone.

Data-driven forecasting of high-dimensional chaotic systems with long short-term memory networks

A data-driven forecasting method for high-dimensional chaotic systems using long short-term memory (LSTM) recurrent neural networks and a hybrid architecture, extending the LSTM with a mean stochastic model (MSM–L STM), is proposed to ensure convergence to the invariant measure.

New perspectives for the prediction and statistical quantification of extreme events in high-dimensional dynamical systems

  • T. Sapsis
  • Physics
    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
  • 2018
This article considers two classes of problems related to extreme events and nonlinear energy transfers, namely the derivation of precursors for the short-term prediction of extreme events, and the efficient sampling of random realizations for the fastest convergence of the probability density function in the tail region.

Predicting observed and hidden extreme events in complex nonlinear dynamical systems with partial observations and short training time series.

A new mathematical framework of building suitable nonlinear approximate models is developed, which aims at predicting both the observed and hidden extreme events in complex nonlinear dynamical systems for short-, medium- and long-range forecasting using only short and partially observed training time series.

Conditional Gaussian Nonlinear System: a Fast Preconditioner and a Cheap Surrogate Model For Complex Nonlinear Systems

This paper aims at exploring the skill of a rich class of nonlinear stochastic models, known as the conditional Gaussian nonlinear system (CGNS), as both a cheap surrogate model and a fast preconditioner for facilitating many computationally challenging tasks.



Nonparametric forecasting of low-dimensional dynamical systems.

The key idea is to represent the discrete shift maps on a smooth basis which can be obtained by the diffusion maps algorithm which converges to a Galerkin projection of the semigroup solution to the underlying dynamics on a basis adapted to the invariant measure.

Analog Forecasting with Dynamics-Adapted Kernels

A suite of forecasting methods which improve traditional analog forecasting by combining ideas from kernel methods developed in harmonic analysis and machine learning and state-space reconstruction for dynamical systems are introduced.

Blended particle filters for large-dimensional chaotic dynamical systems

Test cases for filtering involving the 40-dimensional Lorenz 96 model with a 5-dimensional adaptive subspace for nonlinear blended filtering in various turbulent regimes with at least nine positive Lyapunov exponents demonstrate the high skill of the blended particle filter algorithms in capturing both highly non-Gaussian dynamical features as well as crucial nonlinear statistics for accurate filtering in extreme filtering regimes with sparse infrequent high-quality observations.

Blending Modified Gaussian Closure and Non-Gaussian Reduced Subspace Methods for Turbulent Dynamical Systems

This paper derives a blended framework that links inexpensive second-order uncertainty quantification schemes that model the full space with high order statistical models in specific reduced-order subspaces and allows for the correct prediction of the second- order statistics in theFull space and also the correct modeling of the higher-order statistics in reduced- order subspaced.

Physics constrained nonlinear regression models for time series

A central issue in contemporary science is the development of data driven statistical nonlinear dynamical models for time series of partial observations of nature or a complex physical model. It has

Statistically accurate low-order models for uncertainty quantification in turbulent dynamical systems

  • T. SapsisA. Majda
  • Environmental Science, Physics
    Proceedings of the National Academy of Sciences
  • 2013
A framework for low-order predictive statistical modeling and uncertainty quantification in turbulent dynamical systems is developed here. These reduced-order, modified quasilinear Gaussian (ROMQG)