engleski

Reconstructing State Space from Scalar Time Series with Different Time Delays

Paper presents a technique for a state space reconstruction from chaotic scalar time series. It is based on delay coordinate method, which is extended using different delays when doing state space reconstruction. Compared to the case when equal delays are used, this technique may be used to achieve lower minimum embedding dimension when dealing with poorly sampled time series, but also with some fairly sampled. Alse, it may be applied to select reasonable and reduced set of input variables to neural network when dealing with time series prediction problem.

state space; chaotic time series

nije evidentirano