Search

An improved Fuzzy C-Means (FCM) algorithm, which is called Reliability-based Spatial context Fuzzy C-Means (RSFCM), is proposed for image segmentation in this paper. Aiming to improve the robustness and accuracy of the clustering algorithm, RSFCM integrates neighborhood correlation model with the reliability measurement to describe the spatial relationship of the target. It can make up for the shortcomings of the known FCM algorithm which is sensitive to noise. Furthermore, RSFCM algorithm preserves details of the image by balancing the insensitivity of noise and the reduction of edge blur using a new fuzzy measure indicator. Experimental data consisting of a synthetic image, a brain Magnetic Resonance (MR) image, a remote sensing image, and a traffic sign image are used to test the algorithm's performance. Compared with the traditional fuzzy C-means algorithm, RSFCM algorithm can effectively reduce noise interference, and has better robustness. In comparison with state-of-the-art fuzzy C-means algorithm, RSFCM algorithm could improve pixel separability, suppress heterogeneity of intra-class objects effectively, and it is more suitable for image segmentation. [ABSTRACT FROM AUTHOR]

Due to availability of computational tools for data acquisition, it is very easy to collect many dimensions from an object. Nevertheless, data acquisition from an object in an experiment may have a low number of dimensions. The analysis of low dimensional data has break-through role. But raw and sparse nature of dataset imposes new challenges and requirements for data analysis due to their special and unique characteristics. In the process of overall characterization of low-dimensional data, the data pre-processing plays crucial role. One of the first processes is normalization and standardization process. Therefore, in this paper, I would like to propose novel standardization technique called SDFS (Standardization for Distribution Free Statistics) for nonparametric data analysis. This technique is robust for small sample size with missing values of data points, which commonly exist in real time experiments lead to sparse low-dimensional data. The comprehensive experimental evaluation shows that SDFS standardization is significantly outperforms on existing standardization methods. [ABSTRACT FROM AUTHOR]

Smart, and green cities impose stringent requirements on spectral efficiency (SE), and energy efficiency (EE) of vehicular networks. For the current vehicular ad-hoc networks (VANETs), vehicle's mobility leads to rapid topology changes, and high channel uncertainty. However, clustering schemes for establishing stable clusters, and robust power control (RPC) combating with channel fluctuation are investigated independently. In this paper, joint clustering, and RPC schemes are proposed to optimize the SE, and EE of the involved VANETs. Via the same fixed-length slot, the synchronized interference constraints of cluster heads (CHs) are formed, and offer conditions for RPC. Due to the random channel fluctuations, all CHs’ synchronized interference constraints are formulated as probability constraints. Besides, a pricing-based utility which avoids the separate optimization between SE, and EE is introduced, and the price's impact on the tradeoff between them is involved. Since the probability constraints are intractable, and the unified utility is nonconvex, the Bernstein approximation, and successive convex approximation (SCA) are used to transform the problem into a tractable convex one. Through dual decomposition, two RPC algorithms are proposed to determine the optimal solutions for the fixed price C, and the optimal price C*, respectively. Numerical simulations are used to evaluate the algorithmic performances in high-dynamic system, and the results show that the proposed algorithms are effective. The validity of the clustering method, and the proposed RPC scheme is further verified by comparisons. [ABSTRACT FROM AUTHOR]

We consider the possibilities of using stochastic approximation algorithms with randomization on the input under unknown but bounded interference in studying the clustering of data generated by a mixture of Gaussian distributions. The proposed algorithm, which is robust to external disturbances, allows us to process the data "on the fly" and has a high convergence rate. The operation of the algorithm is illustrated by examples of its use for clustering in various difficult conditions. [ABSTRACT FROM AUTHOR]

Highlights • A new data-dependent method for determining the trimming level in robust clustering. • We monitor the cluster structure as the trimming level changes. • The robust clustering constrains the ratio of eigenvalues of dispersion matrices. • The "car-bike" plot exhibits cluster stability as the eigenvalue constraint varies. • Sizable bibliography relating the work to that in the machine learning and pattern recognition fields. Abstract Monitoring the properties of single sample robust analyses of multivariate data as a function of breakdown point or efficiency leads to the adaptive choice of the best values of these parameters, eliminating arbitrary decisions about their values and so increasing the quality of estimators. Monitoring the trimming proportion in robust cluster analysis likewise leads to improved estimators. We illustrate these procedures on a sample of 424 cows with bovine phlegmon. For clustering we use a method which includes constraints on the eigenvalues of the dispersion matrices, so avoiding thread shaped clusters. The "car-bike" plot reveals the stability of clustering as the trimming level changes. The pattern of clusters and outliers alters appreciably for low levels of trimming. [ABSTRACT FROM AUTHOR]

Highlights • Different from previous studies that use a single label as the gold-standard to decide dialogue breakdown, we focused on inherent subjectivity when people regard dialogue breakdowns and variationality due to the design of conversational agents. • We propose methods for dialogue breakdown detection that first cluster annotators based on their label distributions, and then ensemble detectors trained by different sets of annotators. • Two kinds of neural networks are used as detectors. We assume that breakdowns occur locally (word and phrase levels) and/or globally (sentence and context levels). The convolutional neural network (CNN) is applied to detect former type of breakdowns, and long short-term memory (LSTM) detects the latter type of breakdowns. • Our best model achieves 63.6% F-score on breakdown detection, which is 5.6% higher than the baseline using the conditional random fields. • This is an extended version of our workshop paper submitted to the Dialogue Breakdown Detection Challange 3 (DBDC3). We newly conducted additional experiments that confirm effectiveness of clustering annotators based on their annotation distributions. In addition, detailed error analysis revealed characteristics of the CNN and LSTM on the dialogue breakdown detection task. Abstract Dialogue breakdown is a significant problem in conversational agents. Timely breakdown detection helps the agents quickly recover from mistakes, minimizing the impact on user experience. In this paper, we focus on two problems: variations in determining a response that breakdowns a conversation i.e. , subjectivity, and variations in breakdown types due to designs of conversational agents, i.e. , variationality. To address the subjectivity, which decreases the agreement rate among annotators, our methods detect a dialogue breakdown by ensembling detectors trained by different sets of annotators that are grouped using a clustering algorithm. To address the variationality, our methods apply two types of detector architectures to capture global and local breakdowns. The long short-term memory detector considers the global context and the convolutional neural networks detector is sensitive to the local characteristics. The ensemble of all detectors makes a final judgment. The results of the Japanese task in the Dialogue Breakdown Detection Challenge 3 (DBDC3) confirm that our approach significantly outperforms the baseline, which uses the conventional conditional random field. Detailed error analysis reveals that our encoders based on a convolutional neural network and a long short-term memory have different characteristics. It also confirms the effects of annotator clustering. [ABSTRACT FROM AUTHOR]

In this paper we introduce OpenEnsembles, a Python toolkit for performing and analyzing ensemble clustering. Ensemble clustering is the process of creating many clustering solutions for a given dataset and utilizing the relationships observed across the ensemble to identify final solutions, which are more robust, stable or better than the individual solutions within the ensemble. The OpenEnsembles library provides a unified interface for applying transformations to data, clustering data, visualizing individual clustering solutions, visualizing and finishing the ensemble, and calculating validation metrics for a clustering solution for any given partitioning of the data. We have documented examples of using OpenEnsembles to create, analyze, and visualize a number of different types of ensemble approaches on toy and example datasets. OpenEnsembles is released under the GNU General Public License version 3, can be installed via Conda or the Python Package Index (pip), and is available at https://github.com/NaegleLab/OpenEnsembles. [ABSTRACT FROM AUTHOR]

Highlights • We propose a robust one-shot three-dimensional imaging method based on image processing. • We propose a top to bottom clustering algorithm to cluster the continuous lines robustly. • We propose a phase map computation method based on the clustered lines. Abstract Reconstruction of the object by single-shot is of great importance in many applications, in which the object is moving or its shape is non-rigid and changes irregularly. In this paper, we propose a single-shot structured light 3D imaging technique that calculates the phase map from the distorted line pattern. This technique makes use of the image processing techniques to segment and cluster the projected structured light line pattern from one single captured image. The coordinates of the clustered lines are extracted to form a low-resolution phase matrix which is then transformed to full-resolution phase map by spline interpolation. The 3D shape of the object is computed from the full-resolution phase map and the 2D camera coordinates. Experimental results show that the proposed method was able to reconstruct the three-dimensional shape of the object robustly from one single image. [ABSTRACT FROM AUTHOR]

Necrotic enteritis (NE) is a serious disease of poultry caused by the bacterium C. perfringens. To identify proteins of C. perfringens that confer virulence with respect to NE, the protein secretions of four NE disease-producing strains and one baseline nondisease-producing strain of C. perfringens were examined. The problem then becomes a clustering task, for the identification of two extreme groups of proteins that were produced at either concordantly higher or concordantly lower levels across all four disease-producing strains compared to the baseline, when most of the proteins do not exhibit significant change across all strains. However, the existence of some nuisance proteins of discordant change may severely distort any biologically meaningful cluster pattern. We develop a tailored multivariate clustering approach to robustly identify the proteins of concordant change. Using a three-component normal mixture model as the skeleton, our approach incorporates several constraints to account for biological expectations and data characteristics. More importantly, we adopt a sparse mean-shift parameterization in the reference distribution, coupled with a regularized estimation approach, to flexibly accommodate proteins of discordant change. We explore the connections and differences between our approach and other robust clustering methods, and resolve the issue of unbounded likelihood under an eigenvalue-ratio condition. Simulation studies demonstrate the superior performance of our method compared with a number of alternative approaches. Our protein analysis along with further biological investigations may shed light on the discovery of the complete set of virulence factors in NE. Supplementary materials for this article are available online. [ABSTRACT FROM AUTHOR]

Clustering is a popular topic in data analysis and pattern recognition research. This paper presents an efficient memetic algorithm for clustering tasks. It applies to the whole simulated annealing process rather than those popular methods, only using metropolis criterion, as the local search mechanism to refine the differential evolutionary for improving the accuracy and robustness. The results show that this algorithm performs better than several existing methods in terms of clustering accuracy and efficiency in the majority of the three synthetic and four real life data sets used in this study. Moreover, the presented algorithm is more robust, flexible and not sensitive to the initial value in the unbalanced, overlapped and noisy data sets. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel corpus-based enrichment approach for short text clustering. Since sparseness brings about the problem of insufficient word co-occurrence and lack of context information, previous researches use external sources such as Wikipedia or WordNet to enrich the representation of short text documents, which requires extra resources and might lead to possible inconsistency. On the other hand, corpus-based approaches use no external information in mining short text data. By introducing a set of conjugate definitions to characterize the structures of topics and words, and by proposing a virtual generative procedure for short texts, we perform expansion on short text data. Specifically, new words which may not appear in a short text document were added with a virtual term frequency, and this virtual frequency is obtained from the posterior probabilities of new words given all the words in that document. The complete procedure can be regarded as mapping data points (documents) from the original feature space to a hidden semantic space (topic space). After performing semantic smoothing, data points are then mapped back to the original space. We conduct experiments on two short text datasets, and the results show that the proposed method can effectively address the sparseness problem. For these datasets, our method, using only a basic clustering algorithm, attains a comparable performance with methods based on enrichment with external information sources. [ABSTRACT FROM AUTHOR]

A robust traffic surveillance system is crucial in improving the control and management of traffic systems. Vehicle flow processing primarily involves counting and tracking vehicles; however, due to complex situations such as brightness changes and vehicle partial occlusions, traditional image segmentation methods are unable to segment and count vehicles correctly. This paper presents a novel framework for vision-based vehicle counting and tracking, which consists of four main procedures: foreground detection, feature extraction, feature analysis, and vehicles counting/tracking. Foreground detection intends to generate regions of interest in an image, which are used to produce significant feature points. Vehicles counting and tracking are achieved by analyzing clusters of feature points. As for testing on recorded traffic videos, the proposed framework is verified to be able to separate occluded vehicles and count the number of vehicles accurately and efficiently. By comparing with other methods, we observe that the proposed framework achieves the highest occlusion segment rate and the counting accuracy. [ABSTRACT FROM AUTHOR]

We consider a semi-supervised clustering problem where the locations of the data objects are subject to uncertainty. Each uncertainty set is assumed to be either a closed convex bounded polyhedron or a closed disk. The final clustering is expected to be in accordance with a given number of instance level constraints. The objective function considered minimizes the total of the sum of the violation costs of the unsatisfied instance level constraints and a weighted sum of squared maximum Euclidean distances between the locations of the data objects and the centroids of the clusters they are assigned to. Given a cluster, we first consider the problem of computing its centroid, namely the centroid computation problem under uncertainty (CCPU). We show that the CCPU can be modeled as a second order cone programing problem and hence can be efficiently solved to optimality. As the CCPU is one of the key ingredients of the several algorithms considered in this paper, a subgradient algorithm is also adopted for its faster solution. We then propose a mixed-integer second order cone programing formulation for the considered clustering problem which is only able to solve small-size instances to optimality. For larger instances, approaches from the semi-supervised clustering literature are modified and compared in terms of computational time and quality. [ABSTRACT FROM AUTHOR]

We consider the problem of clustering noisy finite-length observations of stationary ergodic random processes according to their generative models without prior knowledge of the model statistics and the number of generative models. Two algorithms, both using the $L^1$ -distance between estimated power spectral densities (PSDs) as a measure of dissimilarity, are analyzed. The first one, termed nearest neighbor process clustering (NNPC), relies on partitioning the nearest neighbor graph of the observations via spectral clustering. The second algorithm, simply referred to as $k$-means, consists of a single $k$-means iteration with farthest point initialization and was considered before in the literature, albeit with a different dissimilarity measure. We prove that both algorithms succeed with high probability in the presence of noise and missing entries, and even when the generative process PSDs overlap significantly, all provided that the observation length is sufficiently large. Our results quantify the tradeoff between the overlap of the generative process PSDs, the observation length, the fraction of missing entries, and the noise variance. Finally, we provide extensive numerical results for synthetic and real data and find that NNPC outperforms state-of-the-art algorithms in human motion sequence clustering. [ABSTRACT FROM PUBLISHER]

This paper provides a reformulation of the scenario-based two-stage unit commitment problem under uncertainty that allows finding unit-commitment plans that perform reasonably well both in expectation and for the worst case. The proposed reformulation is based on partitioning the sample space of the uncertain factors by clustering the scenarios that approximate their probability distributions. The degree of conservatism of the resulting unit-commitment plan (that is, how close it is to the one provided by a purely robust or stochastic unit-commitment formulation) is controlled by the number of partitions into which the said sample space is split. To efficiently solve the proposed reformulation of the unit-commitment problem under uncertainty, we develop two alternative parallelization and decomposition schemes that rely on a column-and-constraint generation procedure. Finally, we analyze the quality of the solutions provided by this reformulation for a case study based on the IEEE 14-node power system and test the effectiveness of the proposed parallelization and decomposition solution approaches on the larger IEEE 3-Area RTS-96 power system. [ABSTRACT FROM PUBLISHER]

Missing value imputation is a common approach for preprocessing incomplete data sets. In case of data clustering, imputation methods may cause unexpected bias because they may change the underlying structure of the data. In order to avoid prior imputation of missing values the computational operations must be projected on the available data values. In this paper, we apply a robust nan-K-spatmed algorithm to the clustering problem on hyperspectral image data. Robust statistics, such as multivariate medians, are more insensitive to outliers than classical statistics relying on the Gaussian assumptions. They are, however, computationally more intractable due to the lack of closed-form solutions. We will compare robust clustering methods on the bands incomplete data cubes to standard K-means with full data cubes. [ABSTRACT FROM AUTHOR]

Wireless sensor networking is a promising technology that can lead to automatic, intelligent, easier and more secure systems. A wireless sensor network (WSN) consists of small battery powered devices with limited energy resources. One of the major challenges in WSN lies in the energy constraint and computation resources available at the sensor nodes. One way to achieve energy efficiency would be through the use of a clustering technique. In this paper, we propose computational intelligence (CI) approaches to deal with the problem of sensor nodes clustering in a WSN with the ultimate goal to reduce energy expenditures and thus to extend the lifetime of the network. The main motivation is that CI brings about flexibility, autonomous behavior, and robustness against topology changes, communication failures, and scenario changes. The main features of the proposed work span over two aspects. First, four metaheuristics have been adapted to deal with the tackled problem namely genetic algorithms, evolution strategies, particle swarm optimization and artificial bees colony. In this context, a suitable solution representation scheme has been developed and accordingly the algorithms operators, and overall dynamics have been defined. Second, the four developed algorithms are combined with an exhaustive search that is triggered whenever the number of alive sensor nodes drops below a given threshold. The performance of the proposed algorithms has been assessed and compared to the most known state of the art clustering based method namely low energy adaptive clustering hierarchy algorithm (LEACH). The obtained results show that the proposed approaches optimize in an efficient manner the lifetime of WSNs. They also show that the proposed algorithms compete and even outperform LEACH algorithm. [ABSTRACT FROM AUTHOR]

With increasing demand of new wireless applications and increasing number of wireless user's, problem of spectrum scarcity arises. In this context, cognitive radio supports dynamic spectrum access to address spectrum scarcity problem. Cognitive radio defined the cognitive radio nodes by their ability to intelligently adapt the environment to achieve specific objectives through advanced techniques. The variance of channel availability for cognitive radio nodes degrades connectivity and robustness of this type of network; in this case the use of clustering is an effective approach to meet this challenge. Indeed, the geographical areas are homogeneous in terms of type of radio spectrum, radio resources are better allocated by grouping cognitive radio nodes per cluster. Clustering is interesting to effectively manage the spectrum or routing in cognitive radio ad hoc networks. In this paper, we aim to improve connectivity and cooperativeness of cognitive radio nodes based on the improvement of the k-means algorithm. Our proposed algorithm is applied in cognitive radio ad hoc networks. The obtained results in terms of exchange messages and execution time show the feasibility of our algorithm to form clusters in order to improve connectivity and cooperativeness of cognitive radio nodes in the context of cognitive radio ad hoc networks. [ABSTRACT FROM AUTHOR]

This paper proposes and formally evaluates a distributed clustering strategy of swarm of robots into any predefined number of classes. The strategy is an extension of an existing work that is applicable for two classes only [9] . It is implemented and experimentally tested on a swarm of a real swarm of Kilobots. Based only on the local information coming from neighboring robots and the disposition of virtual tokens in the system, the robots of the swarm can be clustered into different classes. The proposed strategy acts in a distributed manner and without need of any global knowledge nor any movement of the robots. Depending on the amount and weight of the tokens available in the system, robots exchange information to reach a token homogeneous disposition. The clustering strategy is inspired by the settling process of liquids of different densities. Using information gathered from neighboring robots, a token density is computed. As a result, the tokens with higher weights cluster first, shifting those of lower weight, until they form differentiated bands for each group, thus completing the clustering process of the robots. [ABSTRACT FROM AUTHOR]

In this paper the Unsupervised Active Learning Method (UALM), a novel clustering method based on the Active Learning Method (ALM) is introduced. ALM is an adaptive recursive fuzzy learning algorithm inspired by some behavioral features of human brain functionality. UALM is a density-based clustering algorithm that relies on discovering densely connected components of data, where it can find clusters of arbitrary shapes. This approach is a noise-robust clustering method. The algorithm first blurs the data points as ink drop patterns, then summarizes the effects of all data points, and finally puts a threshold on the resulting pattern. It uses the connected-component algorithm for finding clusters. Then determines cluster centers by intersecting the narrow-paths. Experimental results confirmed the superiority of our proposed method compared to the two most well-known density-based clustering algorithms, DBSCAN and DENCLUE. [ABSTRACT FROM AUTHOR]

Robust mixture models approaches, which use non-normal distributions have recently been upgraded to accommodate asymmetric data. In this article we propose a new method based on the General Split Gaussian distribution (GSG) and Cross–Entropy Clustering (CEC). The GSG is a flexible density with reasonably small number of parameters which are easy to estimate. We combine the model with a clustering method which allows to treat groups separately and estimate parameters individually in each cluster. Consequently, we introduce an effective clustering algorithm which deals with non-normal data. [ABSTRACT FROM AUTHOR]

With the rapid development of service-oriented computing, cloud computing and big data, a large number of functionally equivalent web services are available on the Internet. Quality of Service (QoS) becomes a differentiating point of services to attract customers. Since the QoS of services varies widely among users due to the unpredicted network, physical location and other objective factors, many Collaborative Filtering based approaches are recently proposed to predict the unknown QoS by employing the historical user-contributed QoS data. However, most existing approaches ignore the data credibility problem and are thus vulnerable to the unreliable QoS data contributed by dishonest users. To address this problem, we propose a trust-aware approach TAP for reliable personalized QoS prediction. Firstly, we cluster the users and calculate the reputation of users based on the clustering information by a beta reputation system. Secondly, a set of trustworthy similar users is identified according to the calculated user reputation and similarity. Finally, we identify a set of similar services by clustering the services and make prediction for active users by combining the QoS data of the trustworthy similar users and similar services. Comprehensive real-world experiments are conducted to demonstrate the effectiveness and robustness of our approach compared with other state-of-the-art approaches. [ABSTRACT FROM AUTHOR]

Feature descriptors have become an increasingly important tool in shape analysis. Features can be extracted and subsequently used to design robust signatures for shape retrieval, correspondence, classification and clustering. In this paper, we present a graph-theoretic framework for 3D shape clustering using the biharmonic distance map and graph regularized sparse coding. While this work focuses primarily on clustering, our approach is fairly general and can be used to tackle other 3D shape analysis problems. In order to seamlessly capture the similarity between feature descriptors, we perform shape clustering on mid-level features that are generated via graph regularized sparse coding. Extensive experiments are carried out on three standard 3D shape benchmarks to demonstrate the much better performance of the proposed clustering approach in comparison with recent state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Due to the fact that limited amounts of labeled data are normally available in real-world, semi-supervised learning has become a popular option, where we expect to use unlabeled data information to improve the learning performance. However, how to use such unlabeled information to make the predicted labels more reliable remains to be a key for any successful learning. In this paper, we propose a semi-supervised learning framework via combination of semi-supervised clustering and semi-supervised classification. In our approach, the predicted labels are selected by both the constrained k -means and safe semi-supervised SVM (S4VMs) to improve the reliability of the predicted labels. Extensive evaluations on collection of benchmarks and real-world action recognition datasets show that the proposed technique outperforms the others. [ABSTRACT FROM AUTHOR]

Business processes constitute an essential asset of organizations while the related process models help to better comprehend the process and therefore to enable effective process analysis or redesign. However, there are several working environments where flows are particularly flexible (e.g., healthcare, customer service) and process models are either very hard to get created, or they fail to reflect reality. The aim of this paper is to support decision-making by providing comprehensible process models in the case of such flexible environments. Following a process mining approach, we propose a methodology to cluster customers’ flows and produce effective summarizations. We propose a novel method to create a similarity metric that is efficient in downgrading the effect of noise and outliers. We use a spectral technique that emphasizes the robustness of the estimated groups, therefore it provides process analysts with clearer process maps. The proposed method is applied to a real case of a healthcare institution delivering valuable insights and showing compelling performance in terms of process models’ complexity and density. [ABSTRACT FROM AUTHOR]

Distributed processing over networks relies on in-network processing and cooperation among neighboring agents. Cooperation is beneficial when agents share a common objective. However, in many applications, agents may belong to different clusters that pursue different objectives. Then, indiscriminate cooperation will lead to undesired results. In this paper, we propose an adaptive clustering and learning scheme that allows agents to learn which neighbors they should cooperate with and which other neighbors they should ignore. In doing so, the resulting algorithm enables the agents to identify their clusters and to attain improved learning and estimation accuracy over networks. We carry out a detailed mean-square analysis and assess the error probabilities of Types I and II, i.e., false alarm and misdetection, for the clustering mechanism. Among other results, we establish that these probabilities decay exponentially with the step-sizes so that the probability of correct clustering can be made arbitrarily close to one. [ABSTRACT FROM PUBLISHER]

A major component of flow cytometry (FCM) data analysis involves gating, which is the process of identifying homogeneous groups of cells. With the rapid development of the portable flow cytometer, manual gating techniques have been unable to meet the demand for accurate and rapid analysis of samples. To provide a practical application for portable devices, we propose a flexible, statistical model-based clustering approach for identifying cell populations in FCM data. This approach, which mimics the manual gating process, employs a finite mixture model with a density function of skew t distribution and estimates parameters via an expectation maximization algorithm. Data analysis from an experiment on a patient's peripheral blood samples have proven that the proposed methodology yields better results in terms of robustness against outliers than current state-of-the-art automated gating methods, has more flexibility in clustering symmetric data and leads to lower misclassification rates (misclassification rates of skew t method is 0.06442) when handling highly asymmetric data. The method we proposed will improve data analysis of portable flow cytometers, especially when the users have no professional training. [ABSTRACT FROM AUTHOR]

Partially missing data sets are a prevailing problem in clustering analysis. We propose a hybrid algorithm combining fuzzy clustering with particle swarm optimization (PSO) for incomplete data clustering, and missing attributes are represented as intervals. Furthermore, we develop a neighbor interval reconstruction (NIR) method based on pre-classification results that estimates the nearest-neighbor interval of missing attribute using the nearest-neighbor rule, which avoids endpoints of intervals determined by different species information, thereby improving the accuracy of missing attribute intervals and enhancing the robustness of missing attribute imputation. Then, the PSO and fuzzy c-means hybrid algorithm are used for clustering the interval-valued data set, and the global optimization ability of the PSO can improve the accuracy of clustering results compared with gradient-based optimization methods. The experimental results for several UCI data sets show the superiority of the proposed NIR hybrid algorithm. [ABSTRACT FROM AUTHOR]

The mean-shift algorithm is a popular algorithm in computer vision and image processing. It can also be cast as a minimum gamma-divergence estimation. In this paper we focus on the 'blurring' mean-shift algorithm, which is one version of the mean-shift process that successively blurs the dataset. The analysis of the blurring mean-shift is relatively more complicated compared to the nonblurring version, yet the algorithm convergence and the estimation consistency have not been well studied in the literature. In this paper we prove both the convergence and the consistency of the blurring mean-shift. We also perform simulation studies to compare the efficiency of the blurring and the nonblurring versions of the mean-shift algorithms. Our results show that the blurring mean-shift has more efficiency. [ABSTRACT FROM AUTHOR]

Some pressure sensor placement methods for leak detection and location in water distribution networks are based on the pressure sensitivity matrix analysis. This matrix depends on the network demands, which are nondeterministic, and the leak magnitudes, which are unknown. In this paper, the robustness of a sensor placement methodology against the fault sensitivity matrix uncertainty is studied. The robustness study is illustrated by means of a small academic network as well as a district metered area in the Barcelona water distribution network. Results reveal that this uncertainty should be taken into account in the sensor placement methodology. [ABSTRACT FROM AUTHOR]

Identifying subgroups of genes from the gene expression of microarray high-dimensionality database is useful in discovering subtypes of cancers in Colon cancer database. Using clustering analysis for identifying cancer types in Colon cancer database is an extremely difficult task because of high-dimensionality gene with noise. Most of the existing clustering methods for colon to achieve types of cancers often hamper the interpretability of the structure. Therefore the aim of this paper is to develop suitable clustering techniques based on fuzzy c-means, the typicality of possibilistic c-means approaches, kernel functions, and neighborhood term to identify similar characters of genes and samples for getting cancer subtypes in the colon cancer database. In order to avoid the random selection of initial prototypes of fuzzy clustering based techniques, this paper presents an algorithm to initialize the cluster prototypes. The performance of proposed methods has been evaluated through experimental work on Synthetic dataset, Wine dataset, IRIS dataset, Checkerboard, Time series, and Thyroid dataset. This paper successfully implements the proposed methods in finding subtypes of cancers in Colon cancer database. Compared with the results of recent existed clustering methods on benchmark datasets and Colon cancer database, this paper has shown that the proposed clustering approach can identify more similar objects of the subgroups than the existed methods. The superiority of the proposed methods has been proved through clustering accuracy. [ABSTRACT FROM AUTHOR]

In this paper, we tackle the problem of estimating the correct number of clusters from a new perspective. The proposed method probabilistically estimates the Average Central Error (ACE), which is the difference between the true cluster centers and their estimations. The novelty of this work is partly in estimating the unavailable ACE by using the available cluster compactness that is the difference between estimated centers and their members. The application of this approach is explored with K-means clustering. The proposed method denoted by Minimum ACE K-means (MACE-means) is shown to have unique advantages both with synthetic and real data. MACE-means clustering is applied to benchmark real world data sets from UCI machine learning repository and other synthesized clusters that represent a wide class of clustering scenarios. Our analysis confirms superiority of MACE-means over the state of the art clustering methods in robustness to the initialization error, accuracy in detecting the correct number of clusters, having less time complexity, and robustness to cluster overlapping. [ABSTRACT FROM AUTHOR]

The consensus clustering technique combines multiple clustering results without accessing the original data. Consensus clustering can be used to improve the robustness of clustering results or to obtain the clustering results from multiple data sources. In this paper, we propose a novel definition of the similarity between points and clusters. With an iterative process, such a definition of similarity can represent how a point should join or leave a cluster clearly, determine the number of clusters automatically, and combine partially overlapping clustering results. We also incorporate the concept of 'clustering fragment' into our method for increased speed. The experimental results show that our algorithm achieves good performances on both artificial data and real data. [ABSTRACT FROM AUTHOR]

Abstract: Applying k-Means to minimize the sum of the intra-cluster variances is the most popular clustering approach. However, after a bad initialization, poor local optima can be easily obtained. To tackle the initialization problem of k-Means, we propose the MinMax k-Means algorithm, a method that assigns weights to the clusters relative to their variance and optimizes a weighted version of the k-Means objective. Weights are learned together with the cluster assignments, through an iterative procedure. The proposed weighting scheme limits the emergence of large variance clusters and allows high quality solutions to be systematically uncovered, irrespective of the initialization. Experiments verify the effectiveness of our approach and its robustness over bad initializations, as it compares favorably to both k-Means and other methods from the literature that consider the k-Means initialization problem. [Copyright &y& Elsevier]

Highlights: [•] A weighted multivariate fuzzy clustering method for interval-valued data is introduced. [•] Weights measure importance of each variable to compute membership degree by cluster. [•] This method is robust to ambiguous cluster membership assignment. [•] An application with interval-valued scientific production data is described. [ABSTRACT FROM AUTHOR]

Nonnegative Matrix Factorization (NMF) has been one of the most widely used clustering techniques for exploratory data analysis. However, since each data point enters the objective function with squared residue error, a few outliers with large errors easily dominate the objective function. In this article, we propose a Robust Manifold Nonnegative Matrix Factorization (RMNMF)method using 2,1-norm and integrating NMF and spectral clustering under the same clustering framework. We also point out the solution uniqueness issue for the existing NMF methods and propose an additional orthonormal constraint to address this problem. With the new constraint, the conventional auxiliary function approach no longer works. We tackle this difficult optimization problem via a novel Augmented Lagrangian Method (ALM)-based algorithm and convert the original constrained optimization problem on one variable into a multivariate constrained problem. The new objective function then can be decomposed into several subproblems that each has a closedform solution. More importantly, we reveal the connection of our method with robust K-means and spectral clustering, and we demonstrate its theoretical significance. Extensive experiments have been conducted on nine benchmark datasets, and all empirical results show the effectiveness of our method. [ABSTRACT FROM AUTHOR]

Reputation systems have contributed much to the success of electronic marketplaces. However, the problem of unfair testimonies has to be addressed effectively to improve the robustness of reputation systems. Until now, most of the existing approaches focus only on reputation systems using binary testimonies, and thus have limited applicability and effectiveness. In this paper, We propose an integrated CLUstering- Based approach called iCLUB to filter unfair testimonies for reputation systems using multinominal testimonies, in an example application of multiagent-based e-commerce. It adopts clustering techniques and considers buyer agents' local as well as global knowledge about seller agents. Experimental evaluation demonstrates the promising results of our approach in filtering various types of unfair testimonies, its robustness against collusion attacks, and better performance compared to competing models. [ABSTRACT FROM AUTHOR]

Abstract: The need for an effective clustering algorithm for Vehicular Ad Hoc Networks (VANETs) is motivated by the recent research in cluster-based MAC and routing schemes. VANETs are highly dynamic and have harsh channel conditions, thus a suitable clustering algorithm must be robust to channel error and must consider node mobility during cluster formation. This work presents a novel, mobility-based clustering scheme for Vehicular Ad hoc Networks, which forms clusters using the Affinity Propagation algorithm in a distributed manner. This proposed algorithm considers node mobility during cluster formation and produces clusters with high stability. Cluster performance was measured in terms of average clusterhead duration, average cluster member duration, average rate of clusterhead change, and average number of clusters. The proposed algorithm is also robust to channel error and exhibits reasonable overhead. Simulation results confirm the superior performance, when compared to other mobility-based clustering techniques. [Copyright &y& Elsevier]

To overcome the disadvantages of the positioning technologies by fuzzy theory in Wireless Sensor Networks (WSNs), positioning algorithms based on information fusion are presented in this article. The fuzzy theory is used to deal with the randomness and fuzziness in the WSNs. And the information fusion is introduced to improve the location accuracy. If the collinearity of the anchor nodes is larger, the misjudged reference nodes may be caused. They are removed by using clustering method. The algorithms in this paper can enhance the location accuracies compared with using the fuzzy theory and alleviate the effect of the RSSI (Received Signal Strength Indication) measure errors. Moreover, the algorithms avoid the high complexity of computation and the requirement of more anchor nodes. Simulation results indicate that the algorithms are more precise, robust as well as with good suitability. [ABSTRACT FROM AUTHOR]

Abstract: Clustering is the process of grouping objects that are similar, where similarity between objects is usually measured by a distance metric. The groups formed by a clustering method are referred as clusters. Clustering is a widely used activity with multiple applications ranging from biology to economics. Each clustering technique has some advantages and disadvantages. Some clustering algorithms may even require input parameters which strongly affect the result. In most cases, it is not possible to choose the best distance metric, the best clustering method, and the best input argument values for an input data set. Therefore, multiple clusterings can be obtained by several distance metrics, several clustering methods, and several input argument values. And, multiple clusterings can be combined into a new and better quality final clustering. We propose a family of combining multiple clustering algorithms that are memory efficient, scalable, robust, and intuitive. Our new algorithms offer tremendous speed gain and low memory requirements by working at cluster level, while producing very good quality final clusters. Extensive experimental evaluations on some very challenging artificially generated and real data sets from a diverse set of domains establish the usefulness of our methods. [Copyright &y& Elsevier]

In 1993, Yang first extended the classification maximum likelihood (CML) to a so-called fuzzy CML, by combining fuzzy c-partitions with the CML function. Fuzzy c-partitions are generally an extension of hard c-partitions. It was claimed that this was more robust. However, the fuzzy CML still lacks some robustness as a clustering algorithm, such as its in-ability to detect different volumes of clusters, its heavy dependence on parameter initializations and the necessity to provide an a priori cluster number. In this paper, we construct a robust fuzzy CML clustering framework that has a robust clustering method. The eigenvalue decomposition of a covariance matrix is firstly considered using the fuzzy CML model. The Bayesian information criterion (BIC) is then used for model selection, in order to choose the best model with the optimal number of clusters. Therefore, the proposed robust fuzzy CML clustering framework exhibits clustering characteristics that are robust in terms of the parameter initialization, robust in terms of the cluster number and also in terms of its capability to detect different volumes of clusters. Numerical examples and real data applications with comparisons are provided, which demonstrate the effectiveness and superiority of the proposed method. [ABSTRACT FROM AUTHOR]

This paper presents a proposal for channel equalization of non-linear time varying communication channels that uses strategies based on quasi type-2 fuzzy sets. Both triangular and trapezoidal quasi type-2 fuzzy equalizers are designed by means of an incremental design approach. A comparative experimental study among several fuzzy equalization architectures is carried out, showing that the equalizer based on trapezoidal quasi type-2 fuzzy sets exhibits the best performance for low and medium levels of uncertainty. The results of this work also suggest that high order fuzzy sets make the equalization process more robust to the variability of the communications channel. [ABSTRACT FROM AUTHOR]

Abstract: Complex wavelet structural similarity (CW-SSIM) index has been recognized as a novel image similarity measure of broad potential applications due to its robustness to small geometric distortions such as translation, scaling and rotation of images. Nevertheless, how to make the best use of it in image classification problems has not been deeply investigated. In this paper, we introduce a series of novel image classification algorithms based on CW-SSIM and use handwritten digit recognition, and face recognition as examples for demonstration. Among the proposed approaches, the best compromise between accuracy and complexity is obtained by the CW-SSIM support vector machine based algorithms, which combines an unsupervised clustering method to divide the training images into clusters with representative images and a supervised learning method based on support vector machines to maximize the classification accuracy. Our experiments show that such a conceptually simple image classification method, which does not involve any registration, intensity normalization or sophisticated feature extraction processes, and does not rely on any modeling of the image patterns or distortion processes, achieves competitive performance with reduced computational cost. [Copyright &y& Elsevier]

This paper studies the problem of thermal control of buildings from the perspective of partitioning them into clusters for decentralized control. A measure of deviation in performance between centralized and decentralized control in the model predictive control framework, referred to as the optimality loss factor, is derived. Another quantity called the fault propagation metric is introduced as an indicator of the robustness of any decentralized architecture to sensing or communication faults. A computationally tractable agglomerative clustering approach is then proposed to determine the decentralized control architectures, which provide a satisfactory trade-off between the underlying optimality and robustness objectives. The potential use of the proposed partitioning methodology is demonstrated using simulated examples. [ABSTRACT FROM AUTHOR]

Wireless sensor networks are composed by large number of small battery powered sensors distributed in an environment. They are responsible to monitor and transmit its physical characteristics. These networks require robust wireless communication protocols that are energy efficient. Thus, it is a challenge for the self organization protocols to provide network survivability and redundancy features. In this paper, we present a novel clustering algorithm called HDED (hybrid distributed, energy-efficient, and dual homed clustering Algorithm), derived from DED (distributed, energy-efficient, and dual homed clustering) which aims some changes on this protocol to increase its performance. Better coverage, energy efficiency, minimum traffic from nodes to base station, balanced energy consumption are the main features of HDED to improve life time of WSN. Simulation results confirm that HDED is effective in prolonging the network lifetime as well as in improving throughput, than DED and EDED. [ABSTRACT FROM AUTHOR]

Abstract: Community detection is a very important problem in social network analysis. Classical clustering approach, -means, has been shown to be very efficient to detect communities in networks. However, -means is quite sensitive to the initial centroids or seeds, especially when it is used to detect communities. To solve this problem, in this study, we propose an efficient algorithm -rank, which selects the top- nodes with the highest rank centrality as the initial seeds, and updates these seeds by using an iterative technique like -means. Then we extend -rank to partition directed, weighted networks, and to detect overlapping communities. The empirical study on synthetic and real networks show that -rank is robust and better than the state-of-the-art algorithms including -means, BGLL, LPA, infomap and OSLOM. [Copyright &y& Elsevier]