Abstract
This paper proposes a mix of noise filtering, fuzzy clustering, neural mapping and predictive techniques for one-subsequence-ahead forecasting of nonstationary time series. Optionally, we may start with de-noising the time series by wavelet decomposition. A non-overlapping subsequence time series clustering procedure with a sliding window is next addressed, by using a lower-bound of the Dynamic Time Warping distance as a dissimilarity measure, when applying the Fuzzy C-Means algorithm. Afterwards, the subsequence time series transition function is learned by neural mapping, consisting of deriving, for each subsequence time series, the degrees to which it belongs to the c cluster prototypes, when the p(c membership degrees of the previous p subsequences are presented as inputs to the neural network. Finally, this transition function is applied to forecasting one-subsequence-ahead time series, as a weighted mean of the c cluster prototypes to which it belongs, and the S&P 500 data are used for testing.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Berndt, D.J., Clifford, J.: Finding patterns in time series: A dynamic programming approach. In: Fayyad, U.M., Piatetsky-Shapiro, G., Smyth, P., Uthurusamy, R. (eds.) Advances in Knowledge Discovery and Data Mining, pp. 229–248. AAAI Press, Menlo Park (1996)
Donoho, D.: Nonlinear Wavelet Methods for Recovery of Signals, Densities and Spectra from Indirect and Noisy Data. In: Daubechies, I. (ed.) Different Perspectives on Wavelets, Proceeding of Symposia in Applied Mathematics, vol. 47, pp. 173–205. Amer. Math. Soc., Providence (1993)
Georgescu, V.: Generalizations of Fuzzy C-Means Algorithm to Granular Feature Spaces, based on Underlying Fuzzy Metrics: Issues and Related Works. In: 13th IFSA World Congress and 6th Conference of EUSFLAT, Lisbon, Portugal, pp. 1791–1796 (2009)
Georgescu, V.: A Time Series Knowledge Mining Framework Exploiting the Synergy between Subsequence Clustering and Predictive Markovian Models. Fuzzy Economic Review XIV(1), 41–66 (2009)
Keogh, E., Pazzani, M.J.: Scaling up dynamic time warping to massive datasets. In: Żytkow, J.M., Rauch, J. (eds.) PKDD 1999. LNCS (LNAI), vol. 1704, pp. 1–11. Springer, Heidelberg (1999)
Keogh, E., Lin, J., Truppel, W.: Clustering of time series subsequences is meaningless: implications for previous and future research. In: 3rd IEEE International Conference on Data Mining, pp. 115–122 (2003)
Keogh, E., Ratanamahatana, C.A.: Exact indexing of dynamic time warping. Knowledge and Information Systems 7, 358–386 (2005)
Mallat, S.G., Peyré, G.: A Wavelet Tour of Signal Processing: The Sparse Way, 3rd edn. Academic Press, London (2009)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Georgescu, V. (2010). A Computational Intelligence Based Framework for One-Subsequence-Ahead Forecasting of Nonstationary Time Series. In: Torra, V., Narukawa, Y., Daumas, M. (eds) Modeling Decisions for Artificial Intelligence. MDAI 2010. Lecture Notes in Computer Science(), vol 6408. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-16292-3_19
Download citation
DOI: https://doi.org/10.1007/978-3-642-16292-3_19
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-16291-6
Online ISBN: 978-3-642-16292-3
eBook Packages: Computer ScienceComputer Science (R0)