• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2004년도 춘계학술대회
• /
• pp.51-60
• /
• 2004

K-means clustering is an iterative algorithm in which items are moved among sets of clusters until the desired set is reached. K-means clustering has been widely used in many applications, such as market research, pattern analysis or recognition, image processing, etc. It can identify dense and sparse regions among data attributes or object attributes. But k-means algorithm requires many hours to get k clusters that we want, because it is more primitive, explorative. In this paper we propose a new method of k-means clustering using a center of gravity for grid-based sample. It is more fast than any traditional clustering method and maintains its accuracy.

• 센서학회지
• /
• 제13권6호
• /
• pp.446-453
• /
• 2004

A portable electronic tongue (E-Tongue) system using an array of ion-selective electrode (ISE) and personal digital assistants (PDA) for recognizing and analyzing food and drink have been designed. By the employment of PDA, the complex algorithm such as fuzzy c-means algorithm (FCMA) could be used in E-Tongue, PUMA could iteratively solve the cluster centers of pre-determined standard patterns. And the membership between the standard patterns and unknown pattern could be analyzed easily by the present E-Tongue combined with PDA.

• Asia pacific journal of information systems
• /
• 제16권1호
• /
• pp.127-144
• /
• 2006

The K-means algorithm is widely used at the initial stage of data analysis in data mining process, partly because of its low time complexity and the simplicity of practical implementation. Cluster validity indices are used along with the algorithm in order to determine the number of clusters as well as the clustering results of datasets. In this paper, we present a performance comparison of sixteen indices, which are selected from forty indices in literature, while considering their applicability to nonhierarchical clustering algorithms. Data sets used in the experiment are generated based on multivariate normal distribution. In particular, four error types including standardization, outlier generation, error perturbation, and noise dimension addition are considered in the comparison. Through the experiment the effects of varying number of points, attributes, and clusters on the performance are analyzed. The result of the simulation experiment shows that Calinski and Harabasz index performs the best through the all datasets and that Davis and Bouldin index becomes a strong competitor as the number of points increases in dataset.

• Journal of Electrical Engineering and Technology
• /
• 제10권4호
• /
• pp.1877-1885
• /
• 2015

This paper presents a new Nearest-Neighbors based weighted representation for images and weighted K-Nearest-Neighbors (WKNN) classifier to improve the precision of image classification using the Bag of Visual Words (BOVW) based models. Scale-invariant feature transform (SIFT) features are firstly extracted from images. Then, the K-means++ algorithm is adopted in place of the conventional K-means algorithm to generate a more effective visual dictionary. Furthermore, the histogram of visual words becomes more expressive by utilizing the proposed weighted vector quantization (WVQ). Finally, WKNN classifier is applied to enhance the properties of the classification task between images in which similar levels of background noise are present. Average precision and absolute change degree are calculated to assess the classification performance and the stability of K-means++ algorithm, respectively. Experimental results on three diverse datasets: Caltech-101, Caltech-256 and PASCAL VOC 2011 show that the proposed WVQ method and WKNN method further improve the performance of classification.

• ETRI Journal
• /
• 제32권4호
• /
• pp.596-602
• /
• 2010

This paper proposes a computationally efficient learning-based super-resolution algorithm using k-means clustering. Conventional learning-based super-resolution requires a huge dictionary for reliable performance, which brings about a tremendous memory cost as well as a burdensome matching computation. In order to overcome this problem, the proposed algorithm significantly reduces the size of the trained dictionary by properly clustering similar patches at the learning phase. Experimental results show that the proposed algorithm provides superior visual quality to the conventional algorithms, while needing much less computational complexity.

• 인터넷정보학회논문지
• /
• 제12권2호
• /
• pp.103-111
• /
• 2011

대규모 데이터에 대한 특성에 따라 몇 개의 클러스터로 군집화하는 클러스터링 기법은 계층적 클러스터링이나 분할 클러스터링 등 다양한 기법이 있는데 그 중에서 K-Means 알고리즘은 구현이 쉬우나 할당-재계산에 소요되는 시간이 증가하게 된다. 본 논문에서는 초기 클러스터 중심들 간의 거리가 최대가 되도록 하여 초기 클러스터 중심들이 고르게 분포되도록 함으로써 할당-재계산 횟수를 줄이고 전체 클러스터링 시간을 감소시키고자 한다.

• 한국정보통신학회논문지
• /
• 제18권3호
• /
• pp.625-630
• /
• 2014

컴퓨터의 발전과 인터넷의 급속한 발전으로 정보의 양이 폭발적으로 증가하게 되었고 이러한 방대한 양의 정보들은 대부분 문서 형태로 관리되기 때문에, 이들을 효과적으로 검색하고 처리하는 방법의 연구가 필요하다. 문서 군집은 문서간의 유사도를 바탕으로 서로 연관된 문서들을 군집화하여 대용량의 문서들을 자동으로 분류하고 검색하고 처리하는데 효율과 정확성을 증대시킨다. 본 논문은 특징 벡터 공간 상의 벡터들로 표현되는 문서들을 K 평균 알고리즘으로 군집화할 때, 주성분 분석을 사용하여 초기 시드점들을 선정함으로써 군집의 효율을 높이는 방법을 제안한다. 실험 결과를 통하여 제안하는 기법이 기존의 K 평균 알고리즘보다 좋은 결과를 얻을 수 있음을 보였다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1988년도 전기.전자공학 학술대회 논문집
• /
• pp.587-590
• /
• 1988

In the case of VLSI computer system control unit design using PLA, optimal state code assignment algorithm to minimize the PLA area is proposed. An optimal state code assignment algorithm which considers output state and logic minimization simultaneously is proposed, and by means of this, algorithm product term is minimized. Also, by means of this algorithm running time and memory capacitance is decreased as against heuristic state code assignment algorithm which uses matrix calculation and considers the constraint relation only. This algorithm is implemented on VAX 11/750 (UNIX4.3BSD). Through the various test example applied proposed algorithm, the efficiency of this algorithm is shown.

• Journal of the Korean Data and Information Science Society
• /
• 제27권3호
• /
• pp.609-619
• /
• 2016

본 논문에서는 대용량 데이터를 처리 및 분석하기 위해 RHadoop 플랫폼에서 실제 데이터와 모의 실험 데이터를 가지고 K-평균 클러스터링을 구현하고, MapReduce의 컴바이너 사용여부에 따른 처리 속도를 비교하고자 한다. 또한, K-평균 클러스터링에서 최적의 군집수 결정방법을 MapReduce 프로그램으로 구현하여 실제 데이터에 적용하고자 한다. 그리고 제안된 RHadoop 플랫폼의 확장 가능성을 보이기 위해 실제 데이터에서 R의 기본 패키지에서 kmeans() 함수와 bigmemory 패키지 상에서 유용한 bigkmeans() 함수와 처리 속도를 비교하고자 한다.

• 한국경영과학회지
• /
• 제23권1호
• /
• pp.1-16
• /
• 1998

The zone clustering problem arising from several area such as deciding the optimal location of ambient measuring stations is to devide the 2-dimensional area into several sub areas in which included individual zone shows simimlar properties. In general, the optimal solution of this problem is very hard to obtain. Therefore, instead of finding an optimal solution, the generation of near optimal solution within the limited time is more meaningful. In this study, the combination of a genetic algorithm and the modified k-means method is used to obtain the near optimal solution. To exploit the genetic algorithm effectively, a representation of chromsomes and appropriate genetic operators are proposed. The k-means method which is originally devised to solve the object clustering problem is modified to improve the solutions obtained from the genetic algorithm. The experiment shows that the proposed method generates the near optimal solution efficiently.