We present a practicable procedure which allows us to decide if a given time series is pure noise, chaotic but distorted by noise, purely chaotic, or a Markov process. This classification is important since the task of modelling and predicting a time series with neural networks is highly related to the knowledge of the memory and the prediction horizon of the process. Our method is based on a measure of the sensitive dependence on the initial conditions which generalizes the information-theoretical concept of Kolmogorov-Sinai entropy.
CITATION STYLE
Schittenkopf, C., & Deco, G. (1996). An information-theoretic measure for the classification of time series. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1112 LNCS, pp. 773–778). Springer Verlag. https://doi.org/10.1007/3-540-61510-5_130
Mendeley helps you to discover research relevant for your work.