Gaining confidence that a clustering algorithm has produced meaningful results and not an accident of its usually heuristic optimization is central to data mining. This is the issue of cluster validity. We propose here a method by which proximity graphs are used to effectively detect border points and measure the margin between clusters. With analysis of boundary situation, we design a framewor...

Cluster analysis finds its place in many applications especially in data analysis, image processing, pattern recognition, market research by grouping customers based on purchasing pattern, classifying documents on web for information discovery, outlier detection applications and act as a tool to gain insight into the distribution of data to observe characteristics of each cluster. This ensures ...

A review of some popular fuzzy cluster validity indices is given. An index that is based on the generalization of silhouettes to fuzzy partitions is compared with the reviewed indices in conjunction with fuzzy c-means clustering.

In this paper we propose a measure of sim­ ilarity /association between two partitions of a set of objects. Our motivation is the desire to use the measure to characterize the quality or accuracy of clustering algorithms by some­ how comparing the clusters they produce with "ground truth" consisting of classes as­ signed by manual means or some other means in whose veracity there is confidence....

We review two clustering algorithms (hard c-means and single linkage) and three indexes of crisp cluster validity (Hubert's statistics, the Davies-Bouldin index, and Dunn's index). We illustrate two deficiencies of Dunn's index which make it overly sensitive to noisy clusters and propose several generalizations of it that are not as brittle to outliers in the clusters. Our numerical examples sh...

Clustering is one of the most well known types of unsupervised learning. Evaluating the quality of results and determining the number of clusters in data is an important issue. Most current validity indices only cover a subset of important aspects of clusters. Moreover, these indices are relevant only for data sets containing at least two clusters. In this paper, a new bounded index for cluster...

Two basic issues for data analysis and kernel-machines design are approached in this paper: determining the number of partitions of a clustering task and the parameters of kernels. A distance metric is presented to determine the similarity between kernels and FCM proximity matrices. It is shown that this measure is maximized, as a function of kernel and FCM parameters, when there is coherence w...

The problem of classifying an image into different homogeneous regions is viewed as the task of clustering the pixels in the intensity space. In particular, medical image segmentation is complex, and automatically detecting regions or clusters of such widely varying sizes is a challenging task. In this paper, we present automatic fuzzy k-means, and kernelized fuzzy c-means algorithms by conside...

This paper presents a cluster validation based document clustering algorithm, which is capable of identifying an important feature subset and the intrinsic value of model order (cluster number). The important feature subset is selected by optimizing a cluster validity criterion subject to some constraint. For achieving model order identification capability, this feature selection procedure is c...

Document clustering aims to automatically group related document into clusters. If two documents are close to each other in the original document space, they are grouped into the same cluster. If the two documents are far away from each other in the original document space, they tend to be grouped into different cluster. The classical clustering algorithms assign each data to exactly one cluste...