κ-均值聚类算法的改进及其在冰脊表面形态分析中的应用

针对传统k-均值聚类算法事先必须获知类别数和难以确定初始聚类中心的缺点,建立了关于聚类中心和类别数k的双层规划模型,结合粒子群算法确定出聚类中心,通过在迭代过程中不断更新准则函数的方法搜索并确定出最佳类别数惫,基于所建模型,提出了一种改进的k-均值聚类算法,并将算法应用于冰脊表面形态分析中.结果表明,算法得到的聚类结果不但具有相邻类别边界清晰的优点,而且能够较好地反映出地理位置和生长环境对冰脊形成的影响.

一类基于贝叶斯信息准则的k均值聚类算法

典型k-均值算法中的聚类数k必须是事先给定的确定值,然而,实际应用中k很难被精确地确定。同时该算法对初始聚类中心的依赖性而导致聚类结果可能陷入局部极小,使得该算法对一些实际问题无效。采用基于密度聚类算法(DBSCAN),在筛选局部代表点时结合贝叶斯信息准则(BIC),得到少量精准反映局部数据分布的BIC核心点。然后,以BIC核心点为初始聚类中心,BIC核心点数量为类别数,对全局数据进行k-均值聚类。实验结果表明,优化的k-均值算法是一种有效可行的聚类算法。

划分系数和总变差相结合的聚类有效性函数

划分系数是聚类有效性检测中常用方法之一 .针对划分系数存在的严重不足 ,本文从一个新的角度对划分系数进行修改 .结合数据的模糊划分得到的总变差 ,提出了二个新的聚类有效性标准 .实验结果表明 ,本文提出的方法具有良好的分类性能 .

Fuzzy c-means text clustering based on topic concept sub-space

To improve the accuracy of text clustering, fuzzy c-means clustering based on topic concept sub-space （TCS2FCM） is introduced for classifying texts. Five evaluation functions are combined to extract key phrases. Concept phrases, as well as the descriptions of final clusters, are presented using WordNet origin from key phrases. Initial centers and membership matrix are the most important factors affecting clustering performance. Orthogonal concept topic sub-spaces are built with the topic concept phrases representing topics of the texts and the initialization of centers and the membership matrix depend on the concept vectors in sub-spaces. The results show that, different from random initialization of traditional fuzzy c-means clustering, the initialization related to text content contributions can improve clustering precision.

ALLIED FUZZY c-MEANS CLUSTERING MODEL

A novel model of fuzzy clustering, i.e. an allied fuzzy c means （AFCM） model is proposed based on the combination of advantages of fuzzy c means （FCM） and possibilistic c means （PCM） clustering. PCM is sensitive to initializations and often generates coincident clusters. AFCM overcomes this shortcoming and it is an ex tension of PCM. Membership and typicality values can be simultaneously produced in AFCM. Experimental re- suits show that noise data can be well processed, coincident clusters are avoided and clustering accuracy is better.

Auto-expanded multi query examples technology in content-based image retrieval

In order to narrow the semantic gap existing in content-based image retrieval （CBIR）,a novel retrieval technology called auto-extended multi query examples （AMQE） is proposed.It expands the single one query image used in traditional image retrieval into multi query examples so as to include more image features related with semantics.Retrieving images for each of the multi query examples and integrating the retrieval results,more relevant images can be obtained.The property of the recall-precision curve of a general retrieval algorithm and the K-means clustering method are used to realize the expansion according to the distance of image features of the initially retrieved images.The experimental results demonstrate that the AMQE technology can greatly improve the recall and precision of the original algorithms.

不平衡数据集异常检测和分类算法

针对传统异常检测模型在面对不平衡样本集时存在参数优化困难、少数类识别效果差等问题,提出一种组合模型以实现不平衡数据集的异常检测和分类。首先利用支撑向量数据描述构造闭合曲面实现“异常”检测,然后提出改进少数样本合成技术对“异常”数据进行过采样以构建平衡数据集,最后利用所提最大类间-类内距K-均值聚类进行自动聚类,实现3种异常数据的分类判决。结果表明,所提方法能够获得较高的异常检测和分类性能,并且具有较强的泛化能力。

C-means-based ant colony algorithm for TSP

To solve the traveling salesman problem with the characteristics of clustering,a novel hybrid algorithm,the ant colony algorithm combined with the C-means algorithm,is presented.In order to improve the speed of convergence,the traveling salesman problem(TSP)data is specially clustered by the C-means algorithm,then,the result is processed by the ant colony algorithm to solve the problem.The proposed algorithm treats the C-means algorithm as a new search operator and adopts a kind of local searching strategy—2-opt,so as to improve the searching performance.Given the cluster number,the algorithm can obtain the preferable solving result.Compared with the three other algorithms—the ant colony algorithm,the genetic algorithm and the simulated annealing algorithm,the proposed algorithm can make the results converge to the global optimum faster and it has higher accuracy.The algorithm can also be extended to solve other correlative clustering combination optimization problems.Experimental results indicate the validity of the proposed algorithm.

低孔渗砂岩孔隙结构分类方法研究

低孔渗砂岩储层孔隙结构复杂,在二维与三维空间中具有较强非均质性,对储集空间的测井响应有极大影响。而压汞曲线、核磁T_2谱、铸体薄片3种资料在一定程度上各自反映了孔隙结构的不同特征,本文分别提取3种资料的微观孔隙结构参数进行相关性分析,最终分别利用主成分分析-模糊均值聚类方法实现孔隙结构分类,其分类结果压汞曲线分形法所得一致。

A K-means clustering based blind multiband spectrum sensing algorithm for cognitive radio

In this paper,a blind multiband spectrum sensing(BMSS)method requiring no knowledge of noise power,primary signal and wireless channel is proposed based on the K-means clustering(KMC).In this approach,the KMC algorithm is used to identify the occupied subband set(OSS)and the idle subband set(ISS),and then the location and number information of the occupied channels are obtained according to the elements in the OSS.Compared with the classical BMSS methods based on the information theoretic criteria(ITC),the new method shows more excellent performance especially in the low signal-to-noise ratio(SNR)and the small sampling number scenarios,and more robust detection performance in noise uncertainty or unequal noise variance applications.Meanwhile,the new method performs more stablely than the ITC-based methods when the occupied subband number increases or the primary signals suffer multi-path fading.Simulation result verifies the effectiveness of the proposed method.

Watershed classification by remote sensing indices： A fuzzy c-means clustering approach

Determining the relatively similar hydrological properties of the watersheds is very crucial in order to readily classify them for management practices such as flood and soil erosion control. This study aimed to identify homogeneous hydrological watersheds using remote sensing data in western Iran. To achieve this goal, remote sensing indices including SAVI, LAI, NDMI, NDVI and snow cover, were extracted from MODIS data over the period 2000 to 2015. Then, a fuzzy method was used to clustering the watersheds based on the extracted indices. A fuzzy c-mean(FCM) algorithm enabled to classify 38 watersheds in three homogeneous groups.The optimal number of clusters was determined through evaluation of partition coefficient, partition entropy function and trial and error. The results indicated three homogeneous regions identified by the fuzzy c-mean clustering and remote sensing product which are consistent with the variations of topography and climate of the study area. Inherently,the grouped watersheds have similar hydrological properties and are likely to need similar management considerations and measures.

Heavy mineral stratigraphy of sediments from the southern outer shelf of the East China Sea since the last glaciation using fuzzy C-means cluster method

Correspondence analysis and fuzzy C-means cluster methods were used to divide the stratigraphy of heavy mineral assemblages, and the sediment sources and depositional dynamics of the environment reconstructed. The assemblages were taken from marine sediments from the late Pleistocene to the Holocene in Core Q43 situated on the outer shelf of the East China Sea. Based on the variable boundaries of the mineral assemblage at 63 and 228 cmbsf （cm below sea floor）, the core might have previously been divided into three sediment strata marked with units Ⅰ, Ⅱ and Ⅲ, which would be consistent with the divided sediment stratum of the core using minor element geochemistry. The downcore distribution of heavy minerals divided the sedimentary sequence into three major units, which were further subdivided into four subunits. The interval between 0 and 63 cmbsf of the core （unit Ⅰ）, which spans the Holocene and the uppermost late Pleistocene, is characterized by a hornblende-epidote-pyroxene assemblage, and contains relatively a smaller amount of schistic mineral and authigenic pyrite. In comparison, the interval between 63 and 228 cmbsf （unit Ⅱ）, is representative of the Last Glacial Maximum （LGM）, and features a hornblende-epidote-magnetite-ilmenite assemblage containing the highest concentrations of heavy minerals and opaque minerals. However, the interval between 228 and 309 cmbsf （unit Ⅲ）, which spans the subinterglacial period, is characterized by a hornblende-authigenic-pyrite-mica assemblage. Relative ratios of some heavy minerals can be used as tracers of clastic sediment sources. The lower part of the sediment core shows the highest magnetite/ilmenite ratio and relatively high hornblende/augite and hornblende/epidote ratios. The middle core shows the highest hornblende/augite and hornblende/epidote ratios, and the lowest magnetite/ilmenite ratio. The upper part exhibits a slightly higher magnetite/ilmenite ratio, and also the lowest hornblende/augite and hornblende/epidote ratios. The distribution of the mineral ratio is consistent with stratigraphic division in heavy mineral data using correspondence analysis and fuzzy C-means clustering. Variations in heavy mineral association and mineral ratio in core Q43 revealed changes in provenance and depositional environment of the southern outer shelf of the East China Sea since the late Pleistocene, well corresponding to interglacial and glacial cycles.

Two-Stage Resource Allocation Scheme for Three-Tier Ultra-Dense Network

In 5 G Ultra-dense Network(UDN), resource allocation is an efficient method to manage inter-small-cell interference. In this paper, a two-stage resource allocation scheme is proposed to supervise interference and resource allocation while establishing a realistic scenario of three-tier heterogeneous network architecture. The scheme consists of two stages: in stage I, a two-level sub-channel allocation algorithm and a power control method based on the logarithmic function are applied to allocate resource for Macrocell and Picocells, guaranteeing the minimum system capacity by considering the power limitation and interference coordination; in stage II, an interference management approach based on K-means clustering is introduced to divide Femtocells into different clusters. Then, a prior sub-channel allocation algorithm is employed for Femtocells in diverse clusters to mitigate the interference and promote system performance. Simulation results show that the proposed scheme contributes to the enhancement of system throughput and spectrum efficiency while ensuring the system energy efficiency.

Development of slope mass rating system using K-means and fuzzy c-means clustering algorithms

Classification systems such as Slope Mass Rating(SMR) are currently being used to undertake slope stability analysis. In SMR classification system, data is allocated to certain classes based on linguistic and experience-based criteria. In order to eliminate linguistic criteria resulted from experience-based judgments and account for uncertainties in determining class boundaries developed by SMR system,the system classification results were corrected using two clustering algorithms, namely K-means and fuzzy c-means(FCM), for the ratings obtained via continuous and discrete functions. By applying clustering algorithms in SMR classification system, no in-advance experience-based judgment was made on the number of extracted classes in this system, and it was only after all steps of the clustering algorithms were accomplished that new classification scheme was proposed for SMR system under different failure modes based on the ratings obtained via continuous and discrete functions. The results of this study showed that, engineers can achieve more reliable and objective evaluations over slope stability by using SMR system based on the ratings calculated via continuous and discrete functions.

Bag-of-visual-words model for artificial pornographic images recognition

It is illegal to spread and transmit pornographic images over internet,either in real or in artificial format.The traditional methods are designed to identify real pornographic images and they are less efficient in dealing with artificial images.Therefore,criminals turn to release artificial pornographic images in some specific scenes,e.g.,in social networks.To efficiently identify artificial pornographic images,a novel bag-of-visual-words based approach is proposed in the work.In the bag-of-words(Bo W)framework,speeded-up robust feature(SURF)is adopted for feature extraction at first,then a visual vocabulary is constructed through K-means clustering and images are represented by an improved Bo W encoding method,and finally the visual words are fed into a learning machine for training and classification.Different from the traditional BoW method,the proposed method sets a weight on each visual word according to the number of features that each cluster contains.Moreover,a non-binary encoding method and cross-matching strategy are utilized to improve the discriminative power of the visual words.Experimental results indicate that the proposed method outperforms the traditional method.

A multi-view K-multiple-means clustering method

The K-multiple-means(KMM)retains the simple and efficient advantages of the K-means algorithm by setting multiple subclasses,and improves its effect on non-convex data sets.And aiming at the problem that it cannot be applied to the Internet on a multi-view data set,a multi-view K-multiple-means(MKMM)clustering method is proposed in this paper.The new algorithm introduces view weight parameter,reserves the design of setting multiple subclasses,makes the number of clusters as constraint and obtains clusters by solving optimization problem.The new algorithm is compared with some popular multi-view clustering algorithms.The effectiveness of the new algorithm is proved through the analysis of the experimental results.

Single Image Super-Resolution by Clustered Sparse Representation and Adaptive Patch Aggregation

A Single Image Super-Resolution (SISR) reconstruction method that uses clustered sparse representation and adaptive patch aggregation is proposed. First, we randomly extract image patch pairs from the training images, and divide these patch pairs into different groups by K-means clustering. Then, we learn an over-complete sub-dictionary pair offline from corresponding group patch pairs. For a given low-resolution patch, we adaptively select one sub-dictionary to reconstruct the high resolution patch online. In addition, non-local self-similarity and steering kernel regression constraints are integrated into patch aggregation to improve the quality of the recovered images. Experiments show that the proposed method is able to realize state-of-the-art performance in terms of both objective evaluation and visual perception.

An algorithm for segmentation of lung ROI by mean-shift clustering combined with multi-scale HESSIAN matrix dot filtering

A new algorithm for segmentation of suspected lung ROI(regions of interest)by mean-shift clustering and multi-scale HESSIAN matrix dot filtering was proposed.Original image was firstly filtered by multi-scale HESSIAN matrix dot filters,round suspected nodular lesions in the image were enhanced,and linear shape regions of the trachea and vascular were suppressed.Then,three types of information,such as,shape filtering value of HESSIAN matrix,gray value,and spatial location,were introduced to feature space.The kernel function of mean-shift clustering was divided into product form of three kinds of kernel functions corresponding to the three feature information.Finally,bandwidths were calculated adaptively to determine the bandwidth of each suspected area,and they were used in mean-shift clustering segmentation.Experimental results show that by the introduction of HESSIAN matrix of dot filtering information to mean-shift clustering,nodular regions can be segmented from blood vessels,trachea,or cross regions connected to the nodule,non-nodular areas can be removed from ROIs properly,and ground glass object(GGO)nodular areas can also be segmented.For the experimental data set of 127 different forms of nodules,the average accuracy of the proposed algorithm is more than 90%.

Research on Image Segmentation Algorithm based on Fuzzy C-mean Clustering

This paper presents a fuzzy C- means clustering image segmentation algorithm based on particle swarm optimization, the method utilizes the strong search ability of particle swarm clustering search center. Because the search clustering center has small amount of calculation according to density, so it can greatly improve the calculation speed of fuzzy C- means algorithm. The experimental results show that, this method can make the fuzzy clustering to obviously improve the speed, so it can achieve fast image segmentation.

Power interconnected system clustering with advanced fuzzy C-mean algorithm

An advanced fuzzy C-mean （FCM） algorithm was proposed for the efficient regional clustering of multi-nodes interconnected systems. Due to various locational prices and regional coherencies for each node and point, modified similarity measure was considered to gather nodes having similar characteristics. The similarity measure was needed to contain locafi0nal prices as well as regional coherency. In order to consider the two properties simultaneously, distance measure of fuzzy C-mean algorithm had to be modified. Regional clustering algorithm for interconnected power systems was designed based on the modified fuzzy C-mean algorithm. The proposed algorithm produces proper classification for the interconnected power system and the results are demonstrated in the example of IEEE 39-bus interconnected electricity system.