Fast global k-means clustering using cluster membership and inequality

In this paper, we present a fast global k-means clustering algorithm by making use of the cluster membership and geometrical information of a data point. This algorithm is referred to as MFGKM. The algorithm uses a set of inequalities developed in this paper to determine a starting point for the jth cluster center of global k-means clustering. Adopting multiple cluster center selection (MCS) for MFGKM, we also develop another clustering algorithm called MFGKM+MCS. MCS determines more than one starting point for each step of cluster split; while the available fast and modified global k-means clustering algorithms select one starting point for each cluster split. Our proposed method MFGKM can obtain the least distortion; while MFGKM+MCS may give the least computing time. Compared to the modified global k-means clustering algorithm, our method MFGKM can reduce the computing time and number of distance calculations by a factor of 3.78-5.55 and 21.13-31.41, respectively, with the average distortion reduction of 5,487 for the Statlog data set. Compared to the fast global k-means clustering algorithm, our method MFGKM+MCS can reduce the computing time by a factor of 5.78-8.70 with the average reduction of distortion of 30,564 using the same data set. The performances of our proposed methods are more remarkable when a data set with higher dimension is divided into more clusters.     

Interval set clustering

Rough k-means clustering describes uncertainty by assigning some objects to more than one cluster. Rough cluster quality index based on decision theory is applicable to the evaluation of rough clustering. In this paper we analyze rough k-means clustering with respect to the selection of the threshold, the value of risk for assigning an object and uncertainty of objects. According to the analysis, clusters presented as interval sets with lower and upper approximations in rough k-means clustering are not adequate to describe clusters. This paper proposes an interval set clustering based on decision theory. Lower and upper approximations in the proposed algorithm are hierarchical and constructed as outer-level approximations and inner-level ones. Uncertainty of objects in out-level upper approximation is described by the assignment of objects among different clusters. Accordingly, ambiguity of objects in inner-level upper approximation is represented by local uniform factors of objects. In addition, interval set clustering can be improved to obtain a satisfactory clustering result with the optimal number of clusters, as well as optimal values of parameters, by taking advantage of the usefulness of rough cluster quality index in the evaluation of clustering. The experimental results on synthetic and standard data demonstrate how to construct clusters with satisfactory lower and upper approximations in the proposed algorithm. The experiments with a promotional campaign for the retail data illustrates the usefulness of interval set clustering for improving rough k-means clustering results.     

Fuzzy clustering algorithms for unsupervised change detection in remote sensing images

In this paper, we propose a context-sensitive technique for unsupervised change detection in multitemporal remote sensing images. The technique is based on fuzzy clustering approach and takes care of spatial correlation between neighboring pixels of the difference image produced by comparing two images acquired on the same geographical area at different times. Since the ranges of pixel values of the difference image belonging to the two clusters (changed and unchanged) generally have overlap, fuzzy clustering techniques seem to be an appropriate and realistic choice to identify them (as we already know from pattern recognition literatures that fuzzy set can handle this type of situation very well). Two fuzzy clustering algorithms, namely fuzzy c-means (FCM) and Gustafson-Kessel clustering (GKC) algorithms have been used for this task in the proposed work. For clustering purpose various image features are extracted using the neighborhood information of pixels. Hybridization of FCM and GKC with two other optimization techniques, genetic algorithm (GA) and simulated annealing (SA), is made to further enhance the performance. To show the effectiveness of the proposed technique, experiments are conducted on two multispectral and multitemporal remote sensing images. A fuzzy cluster validity index (Xie-Beni) is used to quantitatively evaluate the performance. Results are compared with those of existing Markov random field (MRF) and neural network based algorithms and found to be superior. The proposed technique is less time consuming and unlike MRF does not require any a priori knowledge of distributions of changed and unchanged pixels.     

A new algorithm for initial cluster centers in k-means algorithm

Clustering is one of the widely used knowledge discovery techniques to reveal structures in a dataset that can be extremely useful to the analyst. In iterative clustering algorithms the procedure adopted for choosing initial cluster centers is extremely important as it has a direct impact on the formation of final clusters. Since clusters are separated groups in a feature space, it is desirable to select initial centers which are well separated. In this paper, we have proposed an algorithm to compute initial cluster centers for k-means algorithm. The algorithm is applied to several different datasets in different dimension for illustrative purposes. It is observed that the newly proposed algorithm has good performance to obtain the initial cluster centers for the k-means algorithm.     

Fuzzy C-means based clustering for linearly and nonlinearly separable data

In this paper we present a new distance metric that incorporates the distance variation in a cluster to regularize the distance between a data point and the cluster centroid. It is then applied to the conventional fuzzy C-means (FCM) clustering in data space and the kernel fuzzy C-means (KFCM) clustering in a high-dimensional feature space. Experiments on two-dimensional artificial data sets, real data sets from public data libraries and color image segmentation have shown that the proposed FCM and KFCM with the new distance metric generally have better performance on non-spherically distributed data with uneven density for linear and nonlinear separation.     

Unsupervised possibilistic clustering

In fuzzy clustering, the fuzzy c-means (FCM) clustering algorithm is the best known and used method. Since the FCM memberships do not always explain the degrees of belonging for the data well, Krishnapuram and Keller proposed a possibilistic approach to clustering to correct this weakness of FCM. However, the performance of Krishnapuram and Keller's approach depends heavily on the parameters. In this paper, we propose another possibilistic clustering algorithm (PCA) which is based on the FCM objective function, the partition coefficient (PC) and partition entropy (PE) validity indexes. The resulting membership becomes the exponential function, so that it is robust to noise and outliers. The parameters in PCA can be easily handled. Also, the PCA objective function can be considered as a potential function, or a mountain function, so that the prototypes of PCA can be correspondent to the peaks of the estimated function. To validate the clustering results obtained through a PCA, we generalized the validity indexes of FCM. This generalization makes each validity index workable in both fuzzy and possibilistic clustering models. By combining these generalized validity indexes, an unsupervised possibilistic clustering is proposed. Some numerical examples and real data implementation on the basis of the proposed PCA and generalized validity indexes show their effectiveness and accuracy.     

Optimality and stability of the K-hyperline clustering algorithm

K-hyperline clustering is an iterative algorithm based on singular value decomposition and it has been successfully used in sparse component analysis. In this paper, we prove that the algorithm converges to a locally optimal solution for a given set of training data, based on Lloyd’s optimality conditions. Furthermore, the local optimality is shown by developing an Expectation-Maximization procedure for learning dictionaries to be used in sparse representations and by deriving the clustering algorithm as its special case. The cluster centroids obtained from the algorithm are proved to tessellate the space into convex Voronoi regions. The stability of clustering is shown by posing the problem as an empirical risk minimization procedure over a function class. It is proved that, under certain conditions, the cluster centroids learned from two sets of i.i.d. training samples drawn from the same probability space become arbitrarily close to each other, as the number of training samples increase asymptotically.     

Validity index for crisp and fuzzy clusters

In this article, a cluster validity index and its fuzzification is described, which can provide a measure of goodness of clustering on different partitions of a data set. The maximum value of this index, called the PBM-index, across the hierarchy provides the best partitioning. The index is defined as a product of three factors, maximization of which ensures the formation of a small number of compact clusters with large separation between at least two clusters. We have used both the k-means and the expectation maximization algorithms as underlying crisp clustering techniques. For fuzzy clustering, we have utilized the well-known fuzzy c-means algorithm. Results demonstrating the superiority of the PBM-index in appropriately determining the number of clusters, as compared to three other well-known measures, the Davies-Bouldin index, Dunn's index and the Xie-Beni index, are provided for several artificial and real-life data sets.     

On cluster tree for nested and multi-density data clustering

Clustering is one of the important data mining tasks. Nested clusters or clusters of multi-density are very prevalent in data sets. In this paper, we develop a hierarchical clustering approach—a cluster tree to determine such cluster structure and understand hidden information present in data sets of nested clusters or clusters of multi-density. We embed the agglomerative k-means algorithm in the generation of cluster tree to detect such clusters. Experimental results on both synthetic data sets and real data sets are presented to illustrate the effectiveness of the proposed method. Compared with some existing clustering algorithms (DBSCAN, X-means, BIRCH, CURE, NBC, OPTICS, Neural Gas, Tree-SOM, EnDBSAN and LDBSCAN), our proposed cluster tree approach performs better than these methods.     

Validation of overlapping clustering: A random clustering perspective

As a widely used clustering validation measure, the F-measure has received increased attention in the field of information retrieval. In this paper, we reveal that the F-measure can lead to biased views as to results of overlapped clusters when it is used for validating the data with different cluster numbers (incremental effect) or different prior probabilities of relevant documents (prior-probability effect). We propose a new “IMplication Intensity” (IMI) measure which is based on the F-measure and is developed from a random clustering perspective. In addition, we carefully investigate the properties of IMI. Finally, experimental results on real-world data sets show that IMI significantly alleviates biased incremental and prior-probability effects which are inherent to the F-measure.     

Improvement of the k-means clustering filtering algorithm

In this paper, we present a modified filtering algorithm (MFA) by making use of center variations to speed up clustering process. Our method first divides clusters into static and active groups. We use the information of cluster displacements to reject unlikely cluster centers for all nodes in the kd-tree. We reduce the computational complexity of filtering algorithm (FA) through finding candidates for each node mainly from the set of active cluster centers. Two conditions for determining the set of candidate cluster centers for each node from active clusters are developed. Our approach is different from the major available algorithm, which passes no information from one stage of iteration to the next. Theoretical analysis shows that our method can reduce the computational complexity, in terms of the number of distance calculations, of FA at each stage of iteration by a factor of FC/AC, where FC and AC are the numbers of total clusters and active clusters, respectively. Compared with the FA, our algorithm can effectively reduce the computing time and number of distance calculations. It is noted that our proposed algorithm can generate the same clusters as that produced by hard k-means clustering. The superiority of our method is more remarkable when a larger data set with higher dimension is used.     

A top-down approach for density-based clustering using multidimensional indexes

Clustering on large databases has been studied actively as an increasing number of applications involve huge amount of data. In this paper, we propose an efficient top-down approach for density-based clustering, which is based on the density information stored in index nodes of a multidimensional index. We first provide a formal definition of the cluster based on the concept of region contrast partition. Based on this notion, we propose a novel top-down clustering algorithm, which improves the efficiency through branch-and-bound pruning. For this pruning, we present a technique for determining the bounds based on sparse and dense internal regions and formally prove the correctness of the bounds. Experimental results show that the proposed method reduces the elapsed time by up to 96 times compared with that of BIRCH, which is a well-known clustering method. The results also show that the performance improvement becomes more marked as the size of the database increases.     

Mountain c-regressions method

Since Quandt [The estimation of the parameters of a linear regression system obeying two separate regimes, Journal of the American Statistical Association 53 (1958) 873-880] initiated the research on 2-regressions analysis, switching regression had been widely studied and applied in psychology, economics, social science and music perception. In fuzzy clustering, the fuzzy c-means (FCM) is the most commonly used algorithm. Hathaway and Bezdek [Switching regression models and fuzzy clustering, IEEE Transactions on Fuzzy Systems 1 (1993) 195-204] embedded FCM into switching regression where it was called fuzzy c-regressions (FCR). However, the FCR always depends heavily on initial values. In this paper, we propose a mountain c-regressions (MCR) method for solving the initial-value problem. First, we perform data transformation for the switching regression data set, and then implement the modified mountain clustering on the transformed data to extract c cluster centers. These extracted c cluster centers in the transformed space will correspond to c regression models in the original data set. The proposed MCR method can form well-estimated c regression models for switching regression data sets. According to the properties of transformation, the proposed MCR is also robust to noise and outliers. Several examples show the effectiveness and superiority of our proposed method.     

Finding density-based subspace clusters in graphs with feature vectors

Data sources representing attribute information in combination with network information are widely available in today??s applications. To realize the full potential for knowledge extraction, mining techniques like clustering should consider both information types simultaneously. Recent clustering approaches combine subspace clustering with dense subgraph mining to identify groups of objects that are similar in subsets of their attributes as well as densely connected within the network. While those approaches successfully circumvent the problem of full-space clustering, their limited cluster definitions are restricted to clusters of certain shapes. In this work we introduce a density-based cluster definition, which takes into account the attribute similarity in subspaces as well as a local graph density and enables us to detect clusters of arbitrary shape and size. Furthermore, we avoid redundancy in the result by selecting only the most interesting non-redundant clusters. Based on this model, we introduce the clustering algorithm DB-CSC, which uses a fixed point iteration method to efficiently determine the clustering solution. We prove the correctness and complexity of this fixed point iteration analytically. In thorough experiments we demonstrate the strength of DB-CSC in comparison to related approaches.     

Clustering algorithm for intuitionistic fuzzy sets

The intuitionistic fuzzy set (IFS) theory, originated by Atanassov [K. Atanassov, Intuitionistic fuzzy sets, Fuzzy Sets and Systems 20 (1986) 87-96], has been used in a wide range of applications, such as logic programming, medical diagnosis, pattern recognition, and decision making, etc. However, so far there has been little investigation of the clustering techniques of IFSs. In this paper, we define the concepts of association matrix and equivalent association matrix, and introduce some methods for calculating the association coefficients of IFSs. Then, we propose a clustering algorithm for IFSs. The algorithm uses the association coefficients of IFSs to construct an association matrix, and utilizes a procedure to transform it into an equivalent association matrix. The λ-cutting matrix of the equivalent association matrix is used to cluster the given IFSs. Moreover, we extend the algorithm to cluster interval-valued intuitionistic fuzzy sets (IVIFSs), and finally, demonstrate the effectiveness of our clustering algorithm by experimental results.     

Cluster ensembles in collaborative filtering recommendation

Recommender systems, which recommend items of information that are likely to be of interest to the users, and filter out less favored data items, have been developed. Collaborative filtering is a widely used recommendation technique. It is based on the assumption that people who share the same preferences on some items tend to share the same preferences on other items. Clustering techniques are commonly used for collaborative filtering recommendation. While cluster ensembles have been shown to outperform many single clustering techniques in the literature, the performance of cluster ensembles for recommendation has not been fully examined. Thus, the aim of this paper is to assess the applicability of cluster ensembles to collaborative filtering recommendation. In particular, two well-known clustering techniques (self-organizing maps (SOM) and k-means), and three ensemble methods (the cluster-based similarity partitioning algorithm (CSPA), hypergraph partitioning algorithm (HGPA), and majority voting) are used. The experimental results based on the Movielens dataset show that cluster ensembles can provide better recommendation performance than single clustering techniques in terms of recommendation accuracy and precision. In addition, there are no statistically significant differences between either the three SOM ensembles or the three k-means ensembles. Either the SOM or k-means ensembles could be considered in the future as the baseline collaborative filtering technique.     

Some connectivity based cluster validity indices

Identification of the correct number of clusters and the appropriate partitioning technique are some important considerations in clustering where several cluster validity indices, primarily utilizing the Euclidean distance, have been used in the literature. In this paper a new measure of connectivity is incorporated in the definitions of seven cluster validity indices namely, DB-index, Dunn-index, Generalized Dunn-index, PS-index, I-index, XB-index and SV-index, thereby yielding seven new cluster validity indices which are able to automatically detect clusters of any shape, size or convexity as long as they are well-separated. Here connectivity is measured using a novel approach following the concept of relative neighborhood graph. It is empirically established that incorporation of the property of connectivity significantly improves the capabilities of these indices in identifying the appropriate number of clusters. The well-known clustering techniques, single linkage clustering technique and K-means clustering technique are used as the underlying partitioning algorithms. Results on eight artificially generated and three real-life data sets show that connectivity based Dunn-index performs the best as compared to all the other six indices. Comparisons are made with the original versions of these seven cluster validity indices.     

Fuzziness parameter selection in fuzzy c-means: The perspective of cluster validation

Fuzzy c-means （FCM） algorithm is an important clustering method in pattern recognition, while the fuzziness parameter, m, in FCM algorithm is a key parameter that can significantly affect the result of clustering. Cluster validity index （CVI） is a kind of criterion function to validate the clustering results, thereby determining the optimal cluster number of a data set. From the perspective of cluster validation, we propose a novel method to select the optimal value of m in FCM, and four well-known CVIs, namely XB, VK, VT, and SC, for fuzzy clustering are used. In this method, the optimal value of m is determined when CVIs reach their minimum values. Experimental results on four synthetic data sets and four real data sets have demonstrated that the range of m is [2, 3.5] and the optimal interval is [2.5, 3].