An Unsupervised Neural Method for Time Series Analysis, Characterisation and Prediction


We present a novel neural network method for extraction of the embedding function of a time series. We give results on two sets of computer-generated data which are known to show exponentially increasing divergence from nearby initial conditions. We use the network to predict the future evolution of these artificial mappings.


Finite Impulse Response Neural Network Method Artificial Mapping Henon Mapping Residual Output 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    G. Deco and D. Obradovic. An Information Theoretic Approach to Neural Computing. Springer, 1996.Google Scholar
  2. [2]
    C. Robinson. Bifurcation to infinitely many sinks. Communications in Mathematical Physics, pages 433–459, 1990.Google Scholar
  3. [3]
    A. Weigend and N. Gershenfeld. Time Series Prediction, Forecasting the Future and Understanding the Past. Addison Wesley, 1996.Google Scholar

Copyright information

© Springer-Verlag Wien 1998

Authors and Affiliations

  • C. Fyfe
    • 1
  1. 1.Department of Computing and Information SystemsThe University of PaisleyUK

Personalised recommendations