Prediction of Unobserved Bifurcation by Unsupervised Extraction of Slowly Time-Varying System Parameter Dynamics from Time Series Using Reservoir Computing.
Keita TokudaYuichi KatoriPublished in: CoRR (2024)
Keyphrases
- reservoir computing
- recurrent neural networks
- chaotic time series
- prediction accuracy
- financial time series
- dynamical systems
- moving average
- unsupervised learning
- semi supervised
- phase space
- exponential smoothing
- data assimilation
- prediction model
- prediction error
- extreme values
- arma model
- mackey glass
- autoregressive
- sequential data
- automatic extraction
- dynamic time warping
- dynamic model
- feed forward
- parameter space
- non stationary
- anomaly detection
- initial conditions
- prediction algorithm
- unsupervised manner
- learning algorithm
- latent variables
- supervised learning
- wheel slip