融合密度峰值和模糊C-均值聚类算法

针对模糊C—均值(FCM)聚类算法聚类结果依赖于初始中心的选取,易收敛于局部极值等问题,提出了一种密度峰值聚类(DPC)算法和FCM相结合的混合聚类方法(DPC-FCM),利用密度峰值快速搜索算法可以比较准确地刻画聚类初始中心的特点,改善FCM聚类算法存在的不足,从而实现优化聚类.在UCI数据集和人工模拟数据集上的实验结果显示:融合后的新算法和传统的FCM算法相比有着更高的正确率和更快的收敛速度,证明了新算法的可行性.     

An interval number distance- and ranking-based method for remotely sensed image fuzzy clustering

ABSTRACT

Fuzzy c-means clustering is an important non-supervised classification method for remote-sensing images and is based on type-1 fuzzy set theory. Type-1 fuzzy sets use singleton values to express the membership grade; therefore, such sets cannot describe the uncertainty of the membership grade. Interval type-2 fuzzy c-means (IT2FCM) clustering and relevant methods are based on interval type-2 fuzzy sets. Real vectors are used to describe the clustering centres, and the average values of the upper and lower membership grades are used to determine the classification of each pixel. Thus, the width information for interval clustering centres and interval membership grades are ignored. The main contribution of this article is to propose an improved IT2FCM* algorithm by adopting interval number distance (IND) and ranking methods, which use the width information of interval clustering centres and interval membership grades, thus distinguishing this method from existing fuzzy clustering methods. Three different IND definitions are tested, and the distance definition proposed by Li shows the best performance. The second contribution of this work is that two fuzzy cluster validity indices, FS- and XB-, are improved using the IND. Three types of multi/hyperspectral remote-sensing data sets are used to test this algorithm, and the experimental results show that the IT2FCM* algorithm based on the IND proposed by Li performs better than the IT2FCM algorithm using four cluster validity indices, the confusion matrix, and the kappa coefficient (κ). Additionally, the improved FS- index has more indicative ability than the original FS- index.     

一种基于三角模糊数多指标信息的FCM 聚类算法

针对一类具有不确定性三角模糊数多指标信息的聚类分析问题，基于传统的数值信息FCM聚类算法，提出一种新的聚类分析算法．首先描述了具有三角模糊数多指标信息的聚类分析问题，提出并证明了基于三角模糊数多指标信息的关于最优划分和最优聚类中心确定的两个定理；然后根据这两个定理，进一步给出了基于三角模糊数信息的FCM聚类算法的迭代步骤；最后通过一个算例说明了该聚类算法的具体应用．     

A fuzzy self‐organizing map algorithm for biological pattern recognition

Data clustering is a key task for various processes including sequence analysis and pattern recognition. This paper studies a clustering algorithm that aimed to increase accuracy and sensitivity when working with biological data such as DNA sequences. The new algorithm is a modified version of fuzzy C‐means (FCM) and is based on the well‐known self‐organizing map (SOM). In order to show the performance of the algorithm, seven different data sets are processed. The experimental results demonstrate that the proposed algorithm has the potential to outperform SOM and FCM in terms of clustering and classification accuracy abilities. Additionally, a brief comparison is made the proposed algorithm with some previously studied ‘FCM‐SOM’ hybrid algorithms from the literature.     

基于特征加权距离的双指数模糊子空间聚类算法

传统的模糊聚类算法(FCM)使用欧氏距离计算数据点之间的差异时,对于高维数据集聚类效果不够理想.对此,以FCM算法的目标函数为基础,用特征加权距离代替传统的欧氏距离,同时向约束条件中引入指数γ和β,提出了一种基于特征加权距离的双指数模糊子空间聚类算法,并讨论了该算法的收敛性.实验表明,所提出算法可以有效提取高维数据集各类别的相关特征,在真实数据集上有较好的聚类效果.     

基于数据流模型的模糊聚类

模糊聚类是数据挖掘中一个重要聚类算法。当前,基于数据流模型的聚类算法已有了广泛的研究,但这些算法均为硬聚类,尚未见数据流上进行模糊聚类的文献。提出一种针对数据流模型的加权模糊聚类算法,基于真实数据集合和人工数据集的实验表明该算法比传统的模糊聚类算法具有更好的聚类性能。     

基于概率典型性和聚类排斥的无噪声模糊聚类方法

提出了建立在概率典型性和聚类排斥基础上的一个新型无噪声模糊聚类方法RTCM，给出了它的迭代算法过程，并验证了它的收敛性．首先引述了一般的聚类方法，它们主要分为两种：噪声聚类，如模糊c均值（FCM）、可能模糊c均值（FPCM）；无噪声聚类，如NC、PCM等，然后给出了RTCM算法模型和过程，并验证了它的局部收敛性．该算法解决噪声环境下的数据聚类问题，避免了重叠聚类．对比试验表明，该算法改善了噪声环境下FCM，NC、PCM、FPCM的聚类中心质量，有效地解决了PCM在近邻聚类数据中的聚类重叠问题．     

基于熵与邻域约束的模糊C均值改进算法

针对模糊C均值(FCM)聚类算法没有考虑样本不同属性的重要程度、邻域信息等问题,提出一种基于熵与邻域约束的FCM算法。首先通过计算样本各属性的熵值来为各属性赋予权重,结合属性权重改进距离度量函数;随后根据邻域样本与中心样本间的距离计算邻域隶属度权重,加权得到邻域隶属度,利用邻域隶属度约束目标函数,修正隶属度迭代过程,最终达到提升FCM聚类算法性能的目的。理论分析和在人造数据集、多个UCI数据集的试验结果表明,改进后的算法在聚类效果、鲁棒性上均优于传统FCM算法、PCM算法、KFCM算法、KPCM算法和DSFCM算法,表明了本文算法的有效性。     

面向混合属性数据集的改进半监督FCM聚类方法

针对混合属性数据集聚类精度低的问题,本文提出一种基于改进距离度量的半监督模糊均值聚类（Fuzzy C-means,FCM）算法.首先,在数据集中针对类别属性进行预处理,并设置相应的相异度阈值;将传统聚类距离度量与改进的Jaccard距离度量结合,确定混合属性数据集的距离度量函数;最后,将所得距离度量函数与传统半监督FCM算法相结合,并在滚动轴承的不同复合故障数据的特征集中进行聚类.实验表明,该算法能在含无序属性的混合属性数据集的聚类中取得更好的聚类效果.     

A Selection Model for Optimal Fuzzy Clustering Algorithm and Number of Clusters Based on Competitive Comprehensive Fuzzy Evaluation

Fuzzy $c$-means (FCM) and its variants suffer from two problems---local minima and cluster validity---which have a direct impact on the formation of final clustering. There are two strategies---optimization and center initialization strategies---that address the problem of local minima. This paper proposes a center initialization approach based on a minimum spanning tree to keep FCM from local minima. With regard to cluster validity, various strategies have been proposed. On the basis of the fuzzy cluster validity index, this paper proposes a selection model that combines multiple pairs of a fuzzy clustering algorithm and cluster validity index to identify the number of clusters and simultaneously selects the optimal fuzzy clustering for a dataset. The promising performance of the proposed center-initialization method and selection model is demonstrated by experiments on real datasets.      

基于模糊聚类和改进混合蛙跳的协同过滤推荐*

由于传统的协同过滤推荐算法存在很多缺陷,如数据稀疏性、冷启动、低推荐精度等,提出了一种基于模糊聚类和改进混合蛙跳的协同过滤推荐算法。首先利用一种构造的基于时间的指数遗忘函数对原始评分数据进行处理;然后根据得到的基于时间衰退的评分矩阵对用户进行模糊C-均值(FCM)聚类,并找出与目标用户有较高相似性的前几个类作为候选邻居集;再用改进的混合蛙跳算法找到最近邻居集;最后求出目标用户对未参与项目的预测评分。经实验证明,该算法比其他一些算法的推荐精度要高,且由于数据稀疏性引起的不良影响也得到了有效的缓解。     

Fuzzy c-means clustering algorithm for directional data (FCM4DD)

Cluster analysis is a useful tool used commonly in data analysis. The purpose of cluster analysis is to separate data sets into subsets according to their similarities and dissimilarities. In this paper, the fuzzy c-means algorithm was adapted for directional data. In the literature, several methods have been used for the clustering of directional data. Due to the use of trigonometric functions in these methods, clustering is performed by approximate distances. As opposed to other methods, the FCM4DD uses angular difference as the similarity measure. Therefore, the proposed algorithm is a more consistent clustering algorithm than others. The main benefit of FCM4DD is that the proposed method is effectively a distribution-free approach to clustering for directional data. It can be used for N-dimensional data as well as circular data. In addition to this, the importance of the proposed method is that it would be applicable for decision making process, rule-based expert systems and prediction problems. In this study, some existing clustering algorithms and the FCM4DD algorithm were applied to various artificial and real data, and their results were compared. As a result, these comparisons show the superiority of the FCM4DD algorithm in terms of consistency, accuracy and computational time. Fuzzy clustering algorithms for directional data (FCM4DD and FCD) were compared according to membership values and the FCM4DD algorithm obtained more acceptable results than the FCD algorithm.     

基于直方图相关性约束的快速多阈值FCM图像分割算法

针对传统的模糊C均值(FCM)聚类算法在样本数和特征数较多时,运算较为复杂以及耗时较多的问题,本文提出了一种采用直方图的相关性作为约束采样率的快速多阈值FCM分割方法,控制图像失真,使得需要运算的数据量减少,以获得较快的分割速度.由于借助了基于模糊集的图像分割技术--模糊C均值算法实现多阈值图像分割,考虑到了每个像素对...     

一种基于核的快速可能性聚类算法

传统的快速聚类算法大多基于模糊C均值算法（Fuzzy C-means,FCM）,而FCM对初始聚类中心敏感,对噪音数据敏感并且容易收敛到局部极小值,因而聚类准确率不高。可能性C-均值聚类较好地解决了FCM对噪声敏感的问题,但容易产生一致性聚类。将FCM和可能性C-均值聚类结合的聚类算法较好地解决了一致性聚类问题。为进一步提高算法收敛速度和鲁棒性,提出一种基于核的快速可能性聚类算法。该方法引入核聚类的思想,同时使用样本方差对目标函数中参数η进行优化。标准数据集和人造数据集的实验结果表明这种基于核的快速可能性聚类算法提高了算法的聚类准确率,加快了收敛速度。     

新的混合模糊C-均值聚类算法

基于量子行为的粒子群算法(QPSO)是一种改进的粒子群优化算法.它使用的参数个数少,在解的收敛性和全局搜索能力上优于基本的粒子群算法(PSO).将QPSO算法与模糊C-均值(FCM)算法相结合提出一种新的混合模糊C-均值聚类算法(QPSO-FCM),新算法代替了FCM算法的基于梯度下降的迭代过程,在一定程度上克服了FCM算法易陷入局部极小的缺陷,降低了FCM算法的初值敏感度.实验结果表明,改进后的新算法与FCM算法和PSO与FCM结合算法相比,具有良好的收敛性,聚类效果也有较好的改善.     

Density-Weighted Fuzzy c-Means Clustering

In this short paper, a unified framework for performing density-weighted fuzzy $c$-means (FCM) clustering of feature and relational datasets is presented. The proposed approach consists of reducing the original dataset to a smaller one, assigning each selected datum a weight reflecting the number of nearby data, clustering the weighted reduced dataset using a weighted version of the feature or relational data FCM algorithm, and if desired, extending the reduced data results back to the original dataset. Several methods are given for each of the tasks of data subset selection, weight assignment, and extension of the weighted clustering results. The newly proposed weighted version of the non-Euclidean relational FCM algorithm is proved to produce the identical results as its feature data analog for a certain type of relational data. Artificial and real data examples are used to demonstrate and contrast various instances of this general approach.      

Fuzzy order statistics and their application to fuzzy clustering

The median and the median absolute deviation (MAD) are robust statistics based on order statistics. Order statistics are extended to fuzzy sets to define a fuzzy median and a fuzzy MAD. The fuzzy c-means (FCM) clustering algorithm is defined for any p-norm (pFCM), including the l₁-norm (1FCM), The 1FCM clustering algorithm is implemented via the alternating optimization (AO) method and the clustering centers are shown to be the fuzzy median. The resulting AO-1FCM clustering algorithm is called the fuzzy c-medians (FCMED) clustering algorithm. An example illustrates the robustness of the FCMED     

Skin cancer extraction with optimum fuzzy thresholding technique

This paper discusses a new approach to segment different types of skin cancers using fuzzy logic approach. The traditional skin cancer segmentation involves the analysis of image features to delineate the cancerous region from the normal skin. Using low level features such as colour and intensity, segmentation can be done by obtaining a threshold level to separate the two regions. Methods like Otsu optimisation provide a quick and simple process to optimise such threshold level; however this process is prone to the lighting and skin tone variations. Fuzzy clustering algorithm has also been widely used in image processing due to its ability to model the fuzziness of human visual perception. Classical fuzzy C means (FCM) clustering algorithm has been applied to image segmentation with good results; however, the classical FCM is based on type-1 fuzzy sets and is unable to handle uncertainties in the images. In this paper, we proposed an optimum threshold segmentation algorithm based on type-2 fuzzy sets algorithms to delineate the cancerous area from the skin images. By using the 3D colour constancy algorithm, the effect of colour changes and shadows due to skin tone variation in the image can be significantly reduced in the preprocessing stage. We applied the optimum thresholding technique to the preprocessed image over the RGB channels, and combined individual results to achieve the overall skin cancer segmentation. Compared to the Otsu algorithm, the proposed method is less affected by the shadows and skin tone variations. The results also showed more tolerance at the boundary of the cancerous area. Compared with the type-1 FCM algorithm, the proposed method significantly reduced the segmentation error at the normal skin regions.     

基于混合身份搜索黏菌优化的模糊C-均值聚类算法

针对模糊C-均值聚类算法(fuzzy C-means clustering, FCM)对于初始化聚类中心敏感、收敛速度慢,聚类效果不稳定且容易陷入局部最优等问题,提出了一种将黏菌(SMA)与青少年身份搜索(AISA)相融合的自适应优化模糊C-均值算法(AISA-SMA-FCM)。该算法首先通过引入AISA算法中的青少年社会机制,改善SMA算法中的全局搜索和局部开发性能。克服了SMA对于高维数据及部分混峰数据不敏感的缺陷,通过标准测试函数验证改进后的混合AISA-SMA算法寻优求解性能更为优秀;其次此算法用于FCM聚类算法的迭代机制中,通过将AISA-SMA聚类环节加入FCM算法聚类中心迭代过程中,使FCM算法获得自适应优化算法相同的特性,即算法在每次迭代中都将具有探索和开发两个过程,并依据循环迭代次数调节比重,求解聚类结果;最后通过UCI标准数据集仿真测试,利用适应度平均值与聚类正确率评价所提算法的稳定性与有效性,结果表明,AISA-SMA算法用于FCM聚类问题效果较好,AISA-SMA-FCM算法较其他聚类方式和相应的优化技术具有收敛速度快、求解精度高的优点。     

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.