- THESE ARE ONLY PRELIMINARY RESULTS
- To Include
- Data Assimilation Methods
- Narration of results
- Mathematical Methods/Detail
Stochastic Processes Comparison of Accuracy: Data Assimilation, Echo State Machines and Nonlinear Vector Autoregressive Learning Methods.
Comparing discrete and variational data assimilation methods to reservoir computing - machine learning for synthetic chaotic nonlinear dynamical systems
-
The Attractors Used
- Lorenz 1963 Model for Atmospheric Convection (Rayleigh Benard Convection)
- Mackey Glass Dataset
-
Reservoir Computing Methods
- Echo State Machine - Hierarchial Structure
- Echo State Machine - Deep ESN with nonlinear activation function, externally applied
-
Other Neural Network Models
- Nonlinear Vector AutoRegression (NVAR) - Next Generation Reservoir Computer
-
Data Assimilation Methods
- Kalman Filter (Discrete - Ensemble)
- 3D Variational Data Assimilation
- Particle Filter