Repository: Freie Universit├Ąt Berlin, Math Department

On Clustering of Non-stationary Meteorological Time Series

Horenko, I. (2010) On Clustering of Non-stationary Meteorological Time Series. Dyn. of Atm. and Oc., 49 (2-3). pp. 164-187.

[img] PDF - Accepted Version
Available under License Creative Commons Attribution Non-commercial.


Official URL:


A method for clustering of multidimensional non-stationary meteorological time series is presented. The approach is based on optimization of the regularized averaged clustering functional describing the quality of data representation in terms of several regression models and a metastable hidden process switching between them. Proposed numerical clustering algorithm is based on application of the finite element method (FEM) to the problem of non-stationary time series analysis. The main advantage of the presented algorithm compared to Hidden Markov Models (HMMs) and to finite mixture models is that no a priori assumptions about the probability model for the hidden and observed processes (e.g., Markovianity or stationarity) are necessary for the proposed method. Another attractive numerical feature of the discussed algorithm is the possibility to choose the optimal number of metastable clusters and a natural opportunity to control the fuzziness of the resulting decomposition a posteriory, based on the statistical distinguishability of the resulting persistent cluster states. The resulting FEM-K-trends algorithm is compared with some standard fuzzy clustering methods on toy model examples and on analysis of multidimensional historical temperature data locally in Europe and on the global temperature data set.

Item Type:Article
Subjects:Mathematical and Computer Sciences > Mathematics
Divisions:Department of Mathematics and Computer Science > Institute of Mathematics
Department of Mathematics and Computer Science > Institute of Mathematics > BioComputing Group
ID Code:21
Deposited By: Admin Administrator
Deposited On:03 Jan 2009 20:20
Last Modified:16 Nov 2010 09:25

Repository Staff Only: item control page