Nonlinear prediction of chaotic time series

Numerical techniques are presented for constructing nonlinear predictive models directly from time series data and scaling laws are developed which describe the data requirements for reliable predictions.

Abstract

Numerical techniques are presented for constructing nonlinear predictive models directly from time series data. The accuracy of the short-term predictions is tested using computer-generated time series, and comparisons are made of the effectiveness of the various techniques. Scaling laws are developed which describe the data requirements for reliable predictions. It is also shown how to use the models to convincingly distinguish low-dimensional chaos from randomness, and to make statistical long-term predictions.