final_program_abstracts[1]

11 IMSC Session Program
Independent component analysis for extended time series in
climate data
Thursday - Poster Session 6
Fernando Sebastião 1,2 and Irene Oliveria 2,3
1 Department of Mathematics, School of Technology and Management, Polytechnic
Institute of Leiria, Portugal
2 CM-UTAD, Portugal
3 Department of Mathematics, University of Trás-os-Montes and Alto Douro, Portugal
Various techniques of multivariate data analysis have been proposed for sets of time
series, including Multi-channel Singular Spectrum Analysis (MSSA). This technique
is Principal Component Analysis (PCA) (Jolliffe, 2002) of extended matrix of initial
lagged series, hence also designated in the climatological context as Extended
Empirical Orthogonal Function (EEOF) Analysis (von Storch and Zwiers, 1999).
The aim of this work is to present Independent Component Analysis (ICA)
(Hyvärinen et al., 2001) to study the extended matrix of time series, as an alternative
to the method MSSA. ICA is a technique widely used in areas such as image
processing, biomedical signals, telecommunications and econometric time series
among others. In this decade ICA is beginning to be applied in climatology in cases
where the classical PCA does not extract all the essential information underlying a
data set in space and time. Sometimes, ICA is more appropriate than PCA to analyse
time series, since the extraction of Independent Components (ICs) involves higher
order statistics. ICs reveal more useful information than the usual Principal
Components (PCs) since PCA only uses the second order statistics conditioned on the
PCs are no correlated, and which are not necessarily independent. We present an
example of time series for meteorological data and some comparative results between
the techniques under study, particularly with regard to different methods of ordering
ICs, which influence the quality of the reconstructions of the original data.
1. Cheung, Y. and Xu, L. (2001). Independent component ordering in ICA time series
analysis. Neurocomputing, 41, 145-152.
2. Hannachi, A., Jolliffe, I. T. and Stephenson, D. B. (2007). Empirical orthogonal
functions and related techniques in atmospheric science: a review. Internaional
Journal of Climatology, 27, 1119-1152.
3. Hyvärinen, A., Karhunen, J. and Oja, E. (2001). Independent Component Analysis.
John Wiley &Sons, Inc., U.S.A..
4. Jolliffe, I.T. (2002). Principal Component Anlysis (2nd Edition). Springer-Verlag,
New York.
5. von Storch, H. and Zqiers, F. W. (1999). Statistical Analysis in Climate Research.
Sprinfer-Verlag, New York.
Abstracts 287