作者：
Wahlberg, PLantz, GUniv Lund
Dept Elect Engn & Comp Sci Signal Proc Grp S-22100 Lund Sweden Univ Hosp HUG
Dept Neurol Human Brain Mapping Lab CH-1211 Geneva 14 Switzerland
We investigate algorithms for clustering of epileptic electroencephalogram (EEG) spikes. Such a method is useful prior to averaging and inverse computations since the spikes of a patient often belong to a few distinct...
详细信息
We investigate algorithms for clustering of epileptic electroencephalogram (EEG) spikes. Such a method is useful prior to averaging and inverse computations since the spikes of a patient often belong to a few distinct classes. Data sets often contain outliers, which makes algorithms with robust performance desirable. We compare the fuzzyc-means (FcM) algorithm and a graph-theoreticalgorithm. We give criteria for determination of the correct level of outlier contamination. The performance is then studied by aid of simulations, which show good results for a range of circumstances, for both algorithms. The graph-theoretic method gave better results than FcM for simulated signals. Also, when evaluating the methods on seven real-life data sets, the graph-theoretic method was the better method, in terms of closeness to the manual assessment by a neurophysiologist. However, there was some discrepancy between manual and automaticclustering and we suggest as an alternative method a human choice among a limited set of automatically obtained clusterings. Furthermore, we evaluate geometrically weighted feature extraction and conclude that it is useful as a supplementary dimension for clustering.
Large-scale data analysis is a challenging and relevant task for present-day research and industry. As a promising data analysis tool, clustering is becoming more important in the era of big data. In large-scale data ...
详细信息
Large-scale data analysis is a challenging and relevant task for present-day research and industry. As a promising data analysis tool, clustering is becoming more important in the era of big data. In large-scale data clustering, sampling is an efficient and most widely used approximation technique. Recently, several sampling-based clustering algorithms have attracted considerable attention in large-scale data analysis owing to their efficiency. However, some of these existing algorithms have low clustering accuracy, whereas others have high computational complexity. To overcome these deficiencies, a stratified sampling based clustering algorithm for large-scale data is proposed in this paper. Its basic steps include: (1) obtaining a number of representative samples from different strata with a stratified sampling scheme, which are formed by locality sensitive hashing technique, (2) partitioning the chosen samples into different clusters using the fuzzyc-meansclustering algorithm, (3) assigning the out-of-sample objects into their closest clusters via data labeling technique. The performance of the proposed algorithm is compared with the state-of-the-art sampling-based fuzzyc-meansclustering algorithms on several large-scale data sets including synthetic and real ones. The experimental results show that the proposed algorithm outperforms the related algorithms in terms of clustering quality and computational efficiency for large-scale data sets. (c) 2018 Published by Elsevier B.V.
fuzzyc-meansclustering algorithm (FcM), as the most widely used clustering algorithm, works by iteratively updating the membership degree and the cluster centers to improve the effectiveness of clustering. The perfo...
详细信息
fuzzyc-meansclustering algorithm (FcM), as the most widely used clustering algorithm, works by iteratively updating the membership degree and the cluster centers to improve the effectiveness of clustering. The performance of FcM algorithm is chiefly evaluated by intra-cluster compactness and inter-cluster separation. However, it has some defects such as high dependency on the initial cluster centers, sensitivity to the noise samples and outliers, difficulties in obtaining the optimization of hyperparameters, a fairly poor performance on datasets with the nonuniform distribution. The main purpose of this paper is to tackle these issues. The novelty of this paper is three-fold: 1) a new FcM clustering algorithm (i.e., cWAFcM) has been proposed, which has a good capability of performing the clustering on datasets with nonuniform distribution and reducing the high dependency on the initial cluster centers;2) considering the merits of AFcM-SP in removing noise samples and cWAFcM in performing clustering on datasets with nonuniform distribution, a combination of the objective functions of these two clustering algorithms is developed to construct the hybrid AFcM algorithm;and 3) during the parameter setting by means of the PSO algorithm with time-varying acceleration coefficients (PSO-TVAc), a new index, namely adaptive clustering validity index (AcVI), is presented to describe the intra-cluster compactness and the inter-cluster separation in a proper manner. Experiments on six data sets in UcI and one artificial data set have been carried out with a comparison of five well-known FcM algorithms. Experimental results have demonstrated that the proposed hybrid AFcM with adaptive weights can more effectively enhance the performance of FcM to increase the clustering effectiveness than the contrastive algorithms. The ranks for seven algorithms on seven datasets in terms of cVIXB, accuracy, and normalized mutual information(NMI), further verifying the superiority of the new al
Background/purpose: Hyperpigmentation is a common skin problem that looks darker than normal skin regions. Accurate evaluation of a hyperpigmented lesion (HPL) is of clinical importance because proper choice of treatm...
详细信息
Background/purpose: Hyperpigmentation is a common skin problem that looks darker than normal skin regions. Accurate evaluation of a hyperpigmented lesion (HPL) is of clinical importance because proper choice of treatment can be dependent on it. This study aimed to differentiate between epidermal and dermal HPLs. Methods: cross-polarized color images (cPcIs) and fluorescence color images (FcIs) were acquired from the same facial regions. contrast-limited adaptive histogram equalization (cLAHE) was employed to enhance the image contrast and a fuzzy c-means algorithm was implemented to extract the HPLs. The HPLs were superimposed to investigate the difference between cPcI and FcI. Results: The HPL was successfully extracted by applying both cLAHE and fuzzy c-means algorithms. cPcI and FcI resulted in a slightly different HPL, even from the same facial region, indicating a greater percentage area of HPL in FcI than cPcI. conclusion: cPcI and FcI may be utilized to differentiate HPLs that exist in different skin layers. Thus, this approach may contribute to the effective treatment of HPLs.
In this paper a new technique has been proposed for cotton bale management using fuzzy logic. The fuzzyc-meansclustering algorithm has been applied for clustering cotton bales into 5 categories from 1200 randomly ch...
详细信息
In this paper a new technique has been proposed for cotton bale management using fuzzy logic. The fuzzyc-meansclustering algorithm has been applied for clustering cotton bales into 5 categories from 1200 randomly chosen bales of the J-34 variety. In order to cluster bales of different categories, eight fibre properties, viz., the strength, elongation, upper half mean length, length uniformity, short fibre content, micronaire, reflectance and yellowness of each bale have been considered. The fuzzyc-meansclustering method is able to handle the haziness that may be present in the boundaries between adjacent classes of cotton bales as compared to the K-meansclustering method. This method may be used as a convenient tool for the consistent picking of different bale mixes from any number of bales in a warehouse.
In this paper, we psropose a novel method for construction of a distance function and demonstrate its application in image segmentation. In algorithms for image segmentation, distance functions represent a criterion w...
详细信息
In this paper, we psropose a novel method for construction of a distance function and demonstrate its application in image segmentation. In algorithms for image segmentation, distance functions represent a criterion which divides pixels into groups of segments. We introduce two extended aggregation functions, extended powers product and extended weighted arithmetic mean of powers. Their relevant properties are examined, as well as certain resulting properties of distance functions, which are constructed by an application of mentioned aggregation functions. In addition, one pixel descriptor, which is motivated by Local Binary Pattern family of descriptors (LBPs), is introduced and discussed. In the experimental section, we present an application of the introduced extended aggregation functions and descriptor, by a construction of a new distance function, used in fuzzyc-meansclustering algorithm (FcM) for image segmentation. (c) 2019 Elsevier Inc. All rights reserved.
In this paper, a novel variable-rate vector quantizer (VQ) design algorithm using fuzzyclustering technique is presented. The algorithm, termed fuzzy entropy-constrained VQ (FEcVQ) design algorithm, has a better rate...
详细信息
In this paper, a novel variable-rate vector quantizer (VQ) design algorithm using fuzzyclustering technique is presented. The algorithm, termed fuzzy entropy-constrained VQ (FEcVQ) design algorithm, has a better rate-distortion performance than that of the usual entropy-constrained VQ (EcVQ) algorithm for variable-rate VQ design. When performing the fuzzyclustering, the FEcVQ algorithmconsiders both the usual squared-distance measure, and the length of channel index associated with each codeword so that the average rate of the VQ can be controlled. In addition, the membership function for achieving the optimal clustering for the design of FEcVQ are derived. Simulation results demonstrate that the FEcVQ can be an effective alternative for the design of variable-rate VQs.
We propose a design scheme for a hierarchical fuzzy pattern matching classifier (HFPMc) and apply it to the tire tread pattern recognition problem. In this design scheme, a binary decision tree is constructed at first...
详细信息
We propose a design scheme for a hierarchical fuzzy pattern matching classifier (HFPMc) and apply it to the tire tread pattern recognition problem. In this design scheme, a binary decision tree is constructed at first by using fuzzyc-means (FcM) algorithm. At each node, a representative subset of features which can split best the labelled data into two dissimilar groups is selected from all the available features on the base of cluster validity. The cluster validity is evaluated under the two criteria. The one is the polarization degree, and the other is whether all the samples of a class belong to the same cluster or not. Then, a hierarchical cluster structure for the HFPMc is reconstructed by combining the successive nodes formed by the same representative subset of features. As the hierarchical classifier, is used a fuzzy pattern matching classifier in which the designer's intuitive knowledges about the pattern recognition problem can be easily incorporated. At each subhierarchy, the reference fuzzy sets and prototypes for the HFPMc are defined based on the cluster centers of the corresponding subhierarchy. The proposed design scheme is applied to the design of a HFPMc for the tire tread pattern recognition. The design procedure including feature extraction is described in detail. Experimental results show the usefulness of the proposed design scheme.
With the gradual diversification of personalized usage scenarios, user requirements play a direct role in product design decisions. Due to the problem of fuzzy demand caused by user cognitive bias, traditional design ...
详细信息
With the gradual diversification of personalized usage scenarios, user requirements play a direct role in product design decisions. Due to the problem of fuzzy demand caused by user cognitive bias, traditional design methods usually focus on realizing product functions and cannot effectively match user requirements. Therefore, this paper proposes a complex product module division method for user requirements. The method constitutes of three tasks, requirement analysis of module division, design mapping of module division and scheme implementation of module division. Firstly, based on the progressive architecture from initial requirements to precise requirements, the effective user requirements are obtained through similarity recommendation. Secondly, according to the four types of knowledge of function, geometry, physics and design, the design structure matrix is constructed to complete the Requirement-Function-Structure mapping. The improved fuzzy c-means algorithm is used to solve the mapping model, and finally a module division scheme that meets the user requirements is obtained. Taking the chip removal machine as an example, the rationality and effectiveness of the method are verified. The results show that the product modules divided by this method can effectively meet the multiple user requirements.
In multivariate statistical methods, it is important to identify influential observations for a reasonable interpretation of the data structure, In this paper, we propose a method for identifying influential data in t...
详细信息
In multivariate statistical methods, it is important to identify influential observations for a reasonable interpretation of the data structure, In this paper, we propose a method for identifying influential data in the fuzzyc-means (FcM) algorithm, To investigate such data, we consider a perturbation of the data points and evaluate the effect of a perturbation. As a perturbation, we consider two cases: one is the case in which the direction of a perturbation is specified and the other is the case in which the direction of a perturbation is not specified, By computing the change in the clustering result of FcM when given data points are slightly perturbed, we can look for data points that greatly affect the result, Also, we confirm an efficacy of the proposed method by numerical examples.
暂无评论