Nonlinear prediction

  To think about predictability in time series data is worth while even if one is not interested in forecasts at all. Predictability is one way how correlations between data express themselves. These can be linear correlations, nonlinear correlations, or even deterministic contraints. Questions related to those relevant for predictions will reappear with noise reduction and in surrogate data tests, but also for the computation of Lyapunov exponents from data. Prediction is discussed in most of the general nonlinear time series references, in particular, a nice collection of articles can be found in [17].

 
• Model validation
• Simple nonlinear prediction
• Finding unstable periodic orbits
• Locally linear prediction
• Global function fits