Efficient Clustering Method Research Articles

Efficient Cluster Analysis Method for Protein Sequences

This paper proposes an efficient clustering method for protein sequences,using Affinity propagation algorithm(AP) and post-processing.In order to optimize the clustering result,post-processing is used to improve the clustering result of AP.To measure the similarity between two protein sequences,an improved alignment-free similarity measure is presented.This method is evaluated and compared with other algorithms on six protein sequences data sets.Experimental results demonstrate the effective performance of the proposed method.

An Ontology Based Model for Document Clustering

Clustering is an important topic to find relevant content from a document collection and it also reduces the search space. The current clustering research emphasizes the development of a more efficient clustering method without considering the domain knowledge and user’s need. In recent years the semantics of documents have been utilized in document clustering. The discussed work focuses on the clustering model where ontology approach is applied. The major challenge is to use the background knowledge in the similarity measure. This paper presents an ontology based annotation of documents and clustering system. The semi-automatic document annotation and concept weighting scheme is used to create an ontology based knowledge base. The Particle Swarm Optimization (PSO) clustering algorithm can be applied to obtain the clustering solution. The accuracy of clustering has been computed before and after combining ontology with Vector Space Model (VSM). The proposed ontology based framework gives improved performance and better clustering compared to the traditional vector space model. The result using ontology was significant and promising.

SU‐E‐J‐51: Automatic Detection of Markers Based on Threshold Clustering in 3D CT Images for Image‐Guided Radiation Therapy

Purpose: Marker‐based registration, needed especially when there is no sufficient bony anatomy for adequate match, plays an important role in the image guided radiation therapy (IGRT) systems, of which the accuracy of registration greatly depends on that of the location of the markers embed. Since the manual marker detection has many limits for its time consuming and labor intensive, it is essential to enable an automatic detection of makers to help to cut down the human error in registration as well as to get a speedup of the registration process, making IGRT systems more efficient. Methods: In this study, we present a simple and efficient clustering method based on the k‐means clustering algorithm, which we call threshold‐based clustering. This method is established in three procedures. First, we got a set of segmented 3D CT images using binary thresholding method since the value of the markers is far different from that of the CT images. Second, the calculation was done by comparing the spacial distance between every two possible marker candidates to make marker points with similar intensity clustered. Finally, the centers of markers can be obtained after a given number of iteration time which can be changed according to the prior estimated number of markers. Results: This method has been tested on 46 CT slices with a size of 512*512. Markers were successfully identified in at least 99.00%. It took less than a second to detect three markers with the average location accuracy of 0.56mm compared to 1.50mm for the conventional manual technique. Conclusions: It has been testified that the method proposed can accurately detect the location of the markers within 3D CT images. Our results demonstrate the subvoxel accuracy can be achieved completely automatically which makes the method well suited for clinical applications.Supported by the National Natural Science Foundation under grant No.30900386 and the Anhui Provincial Natural Science Foundation under grant No. 090413095 and 11040606Q55

Efficient Clustering Method Based on Rough Set and Genetic Algorithm

In the process of traditional hard clustering, the obtained data objects in clusters are certain. However, the objects in different classes do not have clear boundaries between in reality. A method of dealing with uncertain boundary objects is provided by Rough set theory. Therefore, combing two methods of rough set theory and k-means cluster the objects. At the same time, though the traditional k-means algorithm has powerful local search capability, it easily falls into local optimum. The genetic algorithm can get the global optimal solution, but its convergence is fast. So in the process of clustering, rough set theory and genetic algorithm are introduced. An efficient clustering method based on rough set theory and genetic algorithm is provided. Finally, the experimental results show that the proposed algorithm has the ability to adjust the results and obtain the higher accuracy rate.

CLIC: clustering analysis of large microarray datasets with individual dimension-based clustering

Large microarray data sets have recently become common. However, most available clustering methods do not easily handle large microarray data sets due to their very large computational complexity and memory requirements. Furthermore, typical clustering methods construct oversimplified clusters that ignore subtle but meaningful changes in the expression patterns present in large microarray data sets. It is necessary to develop an efficient clustering method that identifies both absolute expression differences and expression profile patterns in different expression levels for large microarray data sets. This study presents CLIC, which meets the requirements of clustering analysis particularly but not limited to large microarray data sets. CLIC is based on a novel concept in which genes are clustered in individual dimensions first and in which the ordinal labels of clusters in each dimension are then used for further full dimension-wide clustering. CLIC enables iterative sub-clustering into more homogeneous groups and the identification of common expression patterns among the genes separated in different groups due to the large difference in the expression levels. In addition, the computation of clustering is parallelized, the number of clusters is automatically detected, and the functional enrichment for each cluster and pattern is provided. CLIC is freely available at http://gexp2.kaist.ac.kr/clic.

An efficient clustering method for fast rendering of time-varying volumetric medical data

Visualization and exploration of time-varying volumetric medical data help clinicians for better diagnosis and treatment. However, it is a challenge to render these data in an interactive manner because of their complexity and large size. We propose an efficient clustering method for fast compression and rendering of these large dynamic data. We divide the volumes into a set of blocks and use a BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies) based algorithm to cluster them, which can usually find a high quality clustering with a single scan of the blocks. In addition, the granularity of clusters can be adaptively adjusted by dynamically configuring threshold values. In each cluster of blocks, a KeyBlock is generated to represent the cluster, and therefore the storage space of the volumes is reduced greatly. In addition, we assign a lifetime to every KeyBlock and implement a dynamic memory management scheme to further reduce the storage space. During the rendering, each KeyBlock is rendered as a KeyImage, which can be reused if the view transformation and transfer function are not changed. This reuse can help to increase the rendering speed significantly. Experimental results showed that the proposed method can achieve good performance in terms of both speed optimization and space reduction.

An efficient hybrid data clustering method based on K-harmonic means and Particle Swarm Optimization

Clustering is the process of grouping data objects into set of disjoint classes called clusters so that objects within a class are highly similar with one another and dissimilar with the objects in other classes. K-means (KM) algorithm is one of the most popular clustering techniques because it is easy to implement and works fast in most situations. However, it is sensitive to initialization and is easily trapped in local optima. K-harmonic means (KHM) clustering solves the problem of initialization using a built-in boosting function, but it also easily runs into local optima. Particle Swarm Optimization (PSO) algorithm is a stochastic global optimization technique. A hybrid data clustering algorithm based on PSO and KHM (PSOKHM) is proposed in this research, which makes full use of the merits of both algorithms. The PSOKHM algorithm not only helps the KHM clustering escape from local optima but also overcomes the shortcoming of the slow convergence speed of the PSO algorithm. The performance of the PSOKHM algorithm is compared with those of the PSO and the KHM clustering on seven data sets. Experimental results indicate the superiority of the PSOKHM algorithm.

A Graph-Based Biomedical Literature Clustering Approach Utilizing Term's Global and Local Importance Information

In this article, we present a graph-based knowledge representation for biomedical digital library literature clustering. An efficient clustering method is developed to identify the ontology-enriched k-highest density term subgraphs that capture the core semantic relationship information about each document cluster. The distance between each document and the k term graph clusters is calculated. A document is then assigned to the closest term cluster. The extensive experimental results on two PubMed document sets (Disease10 and OHSUMED23) show that our approach is comparable to spherical k-means. The contributions of our approach are the following: (1) we provide two corpus-level graph representations to improve document clustering, a term co-occurrence graph and an abstract-title graph; (2) we develop an efficient and effective document clustering algorithm by identifying k distinguishable class-specific core term subgraphs using terms’ global and local importance information; and (3) the identified term clusters give a meaningful explanation for the document clustering results.

Data privacy protection in multi-party clustering

Privacy concerns over sensitive data have become important in knowledge discovery. Usually, data owners have different levels of concerns over different data attributes, which adds complexity to data privacy. Moreover, collusion among malicious adversaries poses a severe threat to data security. In this paper, we present an efficient clustering method for distributed multi-party data sets using the orthogonal transformation and perturbation techniques. Our method allows data owners to apply different levels of privacy to different attributes. The miner, while receiving the perturbed data, can still obtain accurate clustering results. This method protects data privacy, not only in the semi-honest situation, but also in the presence of collusion. The accuracy of the mining results and the privacy levels, and their relations to the parameters in the method are analyzed. Moreover, we propose an improved version of the method to alleviate the problem with a large number of participants. Experimental results demonstrate the effectiveness of our method as compared to existing methods.

Cluster analysis of massive datasets in astronomy

Clusters of galaxies are a useful proxy to trace the distribution of mass in the universe. By measuring the mass of clusters of galaxies on different scales, one can follow the evolution of the mass distribution (Martinez and Saar, Statistics of the Galaxy Distribution, 2002). It can be shown that finding galaxy clusters is equivalent to finding density contour clusters (Hartigan, Clustering Algorithms, 1975): connected components of the level set S c ?{f>c} where f is a probability density function. Cuevas et al. (Can. J. Stat. 28, 367---382, 2000; Comput. Stat. Data Anal. 36, 441---459, 2001) proposed a nonparametric method for density contour clusters, attempting to find density contour clusters by the minimal spanning tree. While their algorithm is conceptually simple, it requires intensive computations for large datasets. We propose a more efficient clustering method based on their algorithm with the Fast Fourier Transform (FFT). The method is applied to a study of galaxy clustering on large astronomical sky survey data.

Design of cellular manufacturing systems using Latent Semantic Indexing and Self Organizing Maps

A new, efficient clustering method for solving the cellular manufacturing problem is presented in this paper. The method uses the part-machine incidence matrix of the manufacturing system to form machine cells, each of which processes a family of parts. By doing so, the system is decomposed into smaller semi-independent subsystems that are managed more effectively improving overall performance. The proposed method uses Self Organizing Maps (SOMs), a class of unsupervised learning neural networks, to perform direct clustering of machines into cells, without first resorting to grouping parts into families as done by previous approaches. In addition, Latent Semantic Indexing (LSI) is employed to significantly reduce the complexity of the problem resulting in more effective training of the network, significantly improved computational efficiency, and, in many cases, improved solution quality. The robustness of the method and its computational efficiency has been investigated with respect to the dimension of the problem and the degree of dimensionality reduction. The effectiveness of grouping has been evaluated by comparing the results obtained with those of the k-means classical clustering algorithm.

An efficient broadcast data clustering method for multipoint queries in wireless information systems

Mobile computing has become a reality with the convergence of two technologies: powerful portable computers and wireless networks. The restrictions of wireless networks, such as bandwidth and energy limitations make data broadcasting an attractive data communication method. This paper addresses the clustering of wireless broadcast data for multipoint queries. By effective clustering of broadcast data, the mobile client can access the data on the air in short latency. In the paper, we define two affinity measures: data affinity and segment affinity. The data affinity is the degree that two data objects are accessed by queries, and the segment affinity is the degree that two sets of data (i.e., segments) are accessed by queries. Our method clusters data objects based on data and segment affinity measures. We show that the performance of our method is scarcely influenced by the growth of the number of queries.

GMM과 클러스터링 기법에 의한 뉴로-퍼지 시스템 모델링

There have been a lot of considerations dealing with improving the performance of neuro-fuzzy system. The studies on the neuro-fuzzy modeling have largely been devoted to two approaches. First is to improve performance index of system. The other is to reduce the structure size. In spite of its satisfactory result, it should be noted that these are difficult to extend to high dimensional input or to increase the membership functions. We propose a novel neuro-fuzzy system based on the efficient clustering method for initializing the parameters of the premise part. It is a very useful method that maintains a few number of rules and improves the performance. It combine the various algorithms to improve the performance. The Expectation-Maximization algorithm of Gaussian mixture model is an efficient estimation method for unknown parameter estimation of mirture model. The obtained parameters are used for fuzzy clustering method. The proposed method satisfies these two requirements using the Gaussian mixture model and neuro-fuzzy modeling. Experimental results indicate that the proposed method is capable of giving reliable performance.

BIRCH: A New Data Clustering Algorithm and Its Applications

Data clustering is an important technique for exploratory data analysis, and has been studied for several years. It has been shown to be useful in many practical domains such as data classification and image processing. Recently, there has been a growing emphasis on exploratory analysis of very large datasets to discover useful patterns and/or correlations among attributes. This is called data mining, and data clustering is regarded as a particular branch. However existing data clustering methods do not adequately address the problem of processing large datasets with a limited amount of resources (e.g., memory and cpu cycles). So as the dataset size increases, they do not scale up well in terms of memory requirement, running time, and result quality. In this paper, an efficient and scalable data clustering method is proposed, based on a new in-memory data structure called CF-tree, which serves as an in-memory summary of the data distribution. We have implemented it in a system called BIRCH (Balanced Iterative Reducing and Clustering using Hierarchies), and studied its performance extensively in terms of memory requirements, running time, clustering quality, stability and scalability; we also compare it with other available methods. Finally, BIRCH is applied to solve two real-life problems: one is building an iterative and interactive pixel classification tool, and the other is generating the initial codebook for image compression.

An efficient sequential clustering method

A new method for the sequential clustering of data, that gives better results compared to the conventional sequential clustering method, is presented in this paper. This method utilizes certainty factors and it is independent of the technique used for data representation and/or distance computation. Examples are given for cases where the data are represented as feature vectors as well as strings of symbols (syntactic patterns).