A cluster validity index for fuzzy clustering

Zhang, Yunjie; Wang, Weina; Zhang, Xiaona; Yi, Li

doi:10.1016/j.ins.2007.10.004

Cited by 114 publications

(51 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…6, bottom insets) and with the Hellinger distance metrics (F (5,218) ϭ 44.1, p Ͻ 0.005 and F (5,218) ϭ 48.6, p Ͻ 0.001 for RKL and Hellinger metrics, respectively). However, in one dimension, the cluster validity index (Zhang et al, 2008) did not produce a clear determination for the number of clusters (Fig. 7A).…”

Section: Dimensionality Reductionmentioning

confidence: 94%

“…We used instead a cluster validity index proposed by Zhang et al (2008) (Vw). This validity index uses a ratio between a variation measure in each cluster and a separation measure between the fuzzy clusters, and it is reported to have very good performance (Zhang et al, 2008). For the parametric data, we used the fuzzy hyper volume (FHV) and partition density (PD) indices of Gath and Geva (1989), as well.…”

Section: Methodsmentioning

confidence: 99%

“…We did not have access to the data to test correlation of the clusters with the morphological types in the Ruigrok Stats database. This difference between morphology and spontaneous activity may be because spontaneous activity is driven to a large extent by cerebellar inputs and depends on details of the input received by the different neurons or it may be the result of spontaneous activity being driven by intrinsic biophys- (Zhang et al, 2008) does not produce a clear determination for the number of clusters when using the one-dimensional representation in Figure 8C. B, The variation continues to drop as the number of clusters is increased.…”

Section: Morphology Of Clustersmentioning

confidence: 98%

See 2 more Smart Citations

Spontaneous Activity Does Not Predict Morphological Type in Cerebellar Interneurons

et al. 2015

View full text Add to dashboard Cite

The effort to determine morphological and anatomically defined neuronal characteristics from extracellularly recorded physiological signatures has been attempted with varying success in different brain areas. Recent studies have attempted such classification of cerebellar interneurons (CINs) based on statistical measures of spontaneous activity. Previously, such efforts in different brain areas have used supervised clustering methods based on standard parameterizations of spontaneous interspike interval (ISI) histograms. We worried that this might bias researchers toward positive identification results and decided to take a different approach. We recorded CINs from anesthetized cats. We used unsupervised clustering methods applied to a nonparametric representation of the ISI histograms to identify groups of CINs with similar spontaneous activity and then asked how these groups map onto different cell types. Our approach was a fuzzy C-means clustering algorithm applied to the Kullbach-Leibler distances between ISI histograms. We found that there is, in fact, a natural clustering of the spontaneous activity of CINs into six groups but that there was no relationship between this clustering and the standard morphologically defined cell types. These results proved robust when generalization was tested to completely new datasets, including datasets recorded under different anesthesia conditions and in different laboratories and different species (rats). Our results suggest the importance of an unsupervised approach in categorizing neurons according to their extracellular activity. Indeed, a reexamination of such categorization efforts throughout the brain may be necessary. One important open question is that of functional differences of our six spontaneously defined clusters during actual behavior.

show abstract

Section: Dimensionality Reductionmentioning

confidence: 94%

Section: Methodsmentioning

confidence: 99%

Section: Morphology Of Clustersmentioning

confidence: 98%

See 1 more Smart Citation

Spontaneous Activity Does Not Predict Morphological Type in Cerebellar Interneurons

et al. 2015

View full text Add to dashboard Cite

show abstract

“…The literature offers a wide range of evaluation methods for both inter-segmenter comparisons (Borsotti et al 1998;Chabrier et al 2004;Sahoo et al 1988;Weszka & Rosenfeld 1978;Zhang et al 2008a) and cluster quality evaluation (Halkidi et al 2001;Rezaee et al 1998;Zhang et al 2008b). We focused on and adapted here a cluster validity function to be minimized, first proposed by Xie & Beni (1991) (with the parameter m = 2) and generalized by Pal & Bezdek (1995) …”

Section: The Assess Proceduresmentioning

confidence: 99%

Fast and robust segmentation of solar EUV images: algorithm and results for solar cycle 23

Barra¹,

Delouille

Kretzschmar

et al. 2009

A&A

View full text Add to dashboard Cite

Context. The study of the variability of the solar corona and the monitoring of coronal holes, quiet sun and active regions are of great importance in astrophysics as well as for space weather and space climate applications. Aims. In a previous work, we presented the spatial possibilistic clustering algorithm (SPoCA). This is a multi-channel unsupervised spatially-constrained fuzzy clustering method that automatically segments solar extreme ultraviolet (EUV) images into regions of interest. The results we reported on SoHO-EIT images taken from February 1997 to May 2005 were consistent with previous knowledge in terms of both areas and intensity estimations. However, they presented some artifacts due to the method itself. Methods. Herein, we propose a new algorithm, based on SPoCA, that removes these artifacts. We focus on two points: the definition of an optimal clustering with respect to the regions of interest, and the accurate definition of the cluster edges. We moreover propose methodological extensions to this method, and we illustrate these extensions with the automatic tracking of active regions. Results. The much improved algorithm can decompose the whole set of EIT solar images over the 23rd solar cycle into regions that can clearly be identified as quiet sun, coronal hole and active region. The variations of the parameters resulting from the segmentation, i.e. the area, mean intensity, and relative contribution to the solar irradiance, are consistent with previous results and thus validate the decomposition. Furthermore, we find indications for a small variation of the mean intensity of each region in correlation with the solar cycle. Conclusions. The method is generic enough to allow the introduction of other channels or data. New applications are now expected, e.g. related to SDO-AIA data.

show abstract

“…The idea of this algorithm is to achieve the goal of data classification through the membership degree of each sample data belonging to a class. The good clustering result has two characteristics: one is to achieve maximum compactness within cluster; the other is to achieve maximum separation between clusters [3]. However, fuzzy C-means (FCM) algorithm, which is very sensitive to the noise points and the initial center points, is unable to find the global optimum.…”

Section: Introductionmentioning

confidence: 99%

An Improved Fuzzy C-Means Clustering Algorithm Based on Potential Field

Hao¹,

Yu²,

Gao³

et al. 2017

dtcse

View full text Add to dashboard Cite

Abstract. In the area of recommendation system, clustering is an effective way to find similar users and reduce the complexity of recommendation algorithm. A good clustering result has two characteristics: maximizing compactness within cluster and maximizing separation between clusters. Following laws of attraction and repulsion, electric charges in an electric potential field have the same feature as clusters. Based on potential field, an improved Fuzzy C-Means (FCM) clustering algorithm is proposed, which is called EFCM (Electrical FCM). This algorithm combines the basic electrical rules of potential field, Coulomb's law and data field theory to obtain better clustering results. The experiment results show that the improved EFCM algorithm not only obtains good initial cluster centers, but also helps to improve the process of iterative updating.

show abstract

A cluster validity index for fuzzy clustering

Cited by 114 publications

References 43 publications

Spontaneous Activity Does Not Predict Morphological Type in Cerebellar Interneurons

Spontaneous Activity Does Not Predict Morphological Type in Cerebellar Interneurons

Fast and robust segmentation of solar EUV images: algorithm and results for solar cycle 23

An Improved Fuzzy C-Means Clustering Algorithm Based on Potential Field

Contact Info

Product

Resources

About