These tools demonstrate recently developed data-driven nonlinear stochastic-dynamic methods for analysis, modeling and prediction of data from partially-observed systems. 

1. Examples of Empirical Model Reduction [Kondrashov et al, 2005, Kravtsov et al. 2009] within a general class of nonlinear Multilayer Stochastic Models (MSM) with memory effects and complex noise structure  [Kondrashov et al. 2015, Ghil et al. 2018]. 

2. “Past-noise forecasting” example [Chekroun et al. 2011]. 

3. Data-adaptive Harmonic Decomposition [Chekroun and Kondrashov, 2017; Kondrashov et al. 2018] example to  identify coherent spatio-temporal modes in a shorty and noisy dataset.

Download Matlab packages here.