Improved Gath–Geva clustering for fuzzy segmentation of hydrometeorological time series

Abstract

In this paper, an improved Gath-Geva clustering algorithm is proposed for automatic fuzzy segmentation of univariate and multivariate hydrometeorological time series. The algorithm considers time series segmentation problem as Gath-Geva clustering with the minimum message length criterion as segmentation order selection criterion. One characteristic of the improved Gath-Geva clustering algorithm is its unsupervised nature which can automatically determine the optimal segmentation order. Another characteristic is the application of the modified component-wise expectation maximization algorithm in Gath-Geva clustering which can avoid the drawbacks of the classical expectation maximization algorithm: the sensitivity to initialization and the need to avoid the boundary of the parameter space. The other characteristic is the improvement of numerical stability by integrating segmentation order selection into model parameter estimation procedure. The proposed algorithm has been experimentally tested on artificial and hydrometeorological time series. The obtained experimental results show the effectiveness of our proposed algorithm.

Understanding computer usage evolution

“…These approaches are hybrid, as they usually combine a clustering algorithm to detect the segment representatives with maximum likelihood estimation [3], [11]- [ 13]. For instance, the fu zzy maximum likelihood clustering algorithm by Abonyi et al [3] jointly detects segments based on a probabilistic PCA model and fuzzy sets that represent segments in time.…”

Effective intervals determined by information granules to improve forecasting in fuzzy time series

“…In recent years, some methods have been presented for handling forecasting problems based on fuzzy time series such as stock index forecasting , hydrometeorology forecasting (Wang, Liu, & Yin, 2012), enrollment prediction (Song & Chissom, 1993a), temperature forecasting (Wang & Chen, 2009), etc. The concept of fuzzy time series was first proposed by Song and Chissom (1993b).…”

Dynamic programming-based optimization for segmentation and clustering of hydrometeorological time series

“…Abonyi et al (2005) proposed the modified Gath-Geva (GG) clustering algorithm to segment multivariate time series, with fuzzy sets involved in the segmentation process. Wang et al (2012) incorporated a minimum message length criterion into GG clustering to select the order of segmentation.…”

“…For instance, sudden changes in a river flow caused by flood may permanently change the corresponding streamflow (Gedikli et al 2010a). In multivariate time series, comprehensive study of the obtained segmentation results may provide some meaningful insights into the problem under study (Wang et al 2012).…”