• Communications for Statistical Applications and Methods
• /
• v.22 no.1
• /
• pp.55-67
• /
• 2015

An important problem in cluster analysis is the selection of variables that define cluster structure that also eliminate noisy variables that mask cluster structure; in addition, outlier detection is a fundamental task for cluster analysis. Here we provide an automated K-means clustering process combined with variable selection and outlier identification. The Automated K-means clustering procedure consists of three processes: (i) automatically calculating the cluster number and initial cluster center whenever a new variable is added, (ii) identifying outliers for each cluster depending on used variables, (iii) selecting variables defining cluster structure in a forward manner. To select variables, we applied VS-KM (variable-selection heuristic for K-means clustering) procedure (Brusco and Cradit, 2001). To identify outliers, we used a hybrid approach combining a clustering based approach and distance based approach. Simulation results indicate that the proposed automated K-means clustering procedure is effective to select variables and identify outliers. The implemented R program can be obtained at http://www.knou.ac.kr/~sskim/SVOKmeans.r.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.3
• /
• pp.976-990
• /
• 2020

The k-means algorithm is widely used in academia and industry due to easy and simple implementation, enabling fast learning for complex datasets. However, k-means struggles to classify datasets without prior knowledge of specific domains. We proposed the repulsive k-means (RK-means) algorithm in a previous study to improve the k-means algorithm, using the repulsive force concept, which allows deleting unnecessary cluster centroids. Accordingly, the RK-means enables to classifying of a dataset without domain knowledge. However, three main problems remain. The RK-means algorithm includes a cluster repulsive force offset, for clusters confined in other clusters, which can cause cluster locking; we were unable to prove RK-means provided optimal convergence in the previous study; and RK-means shown better performance only normalize term and weight. Therefore, this paper proposes the advanced RK-means (ARK-means) algorithm to resolve the RK-means problems. We establish an initialization strategy for deploying cluster centroids and define a metric for the ARK-means algorithm. Finally, we redefine the mass and normalize terms to close to the general dataset. We show ARK-means feasibility experimentally using blob and iris datasets. Experiment results verify the proposed ARK-means algorithm provides better performance than k-means, k'-means, and RK-means.

• The Korean Journal of Applied Statistics
• /
• v.25 no.3
• /
• pp.471-483
• /
• 2012

One of the most important problems in cluster analysis is the selection of variables that truly define cluster structure, while eliminating noisy variables that mask such structure. Brusco and Cradit (2001) present VS-KM(variable-selection heuristic for K-means clustering) procedure for selecting true variables for K-means clustering based on adjusted Rand index. This procedure starts with the fixed number of clusters in K-means and adds variables sequentially based on an adjusted Rand index. This paper presents an updated procedure combining the VS-KM with the automated K-means procedure provided by Kim (2009). This automated variable selection procedure for K-means clustering calculates the cluster number and initial cluster center whenever new variable is added and adds a variable based on adjusted Rand index. Simulation result indicates that the proposed procedure is very effective at selecting true variables and at eliminating noisy variables. Implemented program using R can be obtained on the website "http://faculty.knou.ac.kr/sskim/nvarkm.r and vnvarkm.r".

• The Korean Journal of Applied Statistics
• /
• v.34 no.6
• /
• pp.905-922
• /
• 2021

Cluster analysis is one of unsupervised learning techniques used for discovering clusters when there is no prior knowledge of group membership. K-means, one of the commonly used cluster analysis techniques, may fail when the number of variables becomes large. In such high-dimensional cases, it is common to perform tandem analysis, K-means cluster analysis after reducing the number of variables using dimension reduction methods. However, there is no guarantee that the reduced dimension reveals the cluster structure properly. Principal component analysis may mask the structure of clusters, especially when there are large variances for variables that are not related to cluster structure. To overcome this, techniques that perform dimension reduction and cluster analysis simultaneously have been suggested. This study proposes probabilistic reduced K-means, the transition of reduced K-means (De Soete and Caroll, 1994) into a probabilistic framework. Simulation shows that the proposed method performs better than tandem clustering or clustering without any dimension reduction. When the number of the variables is larger than the number of samples in each cluster, probabilistic reduced K-means show better formation of clusters than non-probabilistic reduced K-means. In the application to a real data set, it revealed similar or better cluster structure compared to other methods.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.447-449
• /
• 2021

Various wireless sensor network protocols have been proposed to maintain the network for a long time by minimizing energy consumption. Using the K-means clustering algorithm takes longer to cluster than traditional hierarchical algorithms because the center point must be moved repeatedly until the final cluster is established. For K-means clustering-based protocols, only the residual energy of nodes or nodes near the center point of the cluster is considered when the cluster head is elected. In this paper, we propose a new wireless sensor network protocol based on K-means clustering to improve the energy efficiency while improving the aforementioned problems.

• Journal of Korea Multimedia Society
• /
• v.17 no.10
• /
• pp.1160-1170
• /
• 2014

Clustering methods are very useful in many fields such as data mining, classification, and object recognition. Both the supervised and unsupervised grouping approaches can classify a series of sample data with a predefined or automatically assigned cluster number. However, there is no constraint on the number of elements for each cluster. Numbers of cluster members for each cluster obtained from clustering schemes are usually random. Thus, some clusters possess a large number of elements whereas others only have a few members. In some areas such as logistics management, a fixed number of members are preferred for each cluster or logistic center. Consequently, it is necessary to design a clustering method that can automatically adjust the number of group elements. In this paper, a k-means based clustering method with a fixed number of cluster members is proposed. In the proposed method, first, the data samples are clustered using the k-means algorithm. Then, the number of group elements is adjusted by employing a greedy strategy. Experimental results demonstrate that the proposed clustering scheme can classify data samples efficiently for a fixed number of cluster members.

• Journal of the Korean Society for information Management
• /
• v.21 no.4 s.54
• /
• pp.173-185
• /
• 2004

Fast and high-quality document clustering algorithms play an important role in providing data exploration by organizing large amounts of information into a small number of meaningful clusters. Many papers have shown that the hierarchical clustering method takes good-performance, but is limited because of its quadratic time complexity. In contrast, with a large number of variables, K-means has a time complexity that is linear in the number of documents, but is thought to produce inferior clusters. In this paper, Condor system using K-Means algorithm Compares with regular method that the initial centroids have been established in advance, our method performance has been improved a lot.

• The KIPS Transactions:PartB
• /
• v.11B no.7 s.96
• /
• pp.867-874
• /
• 2004

K-Means algorithm is a non-hierarchical (plat) and reassignment techniques and iterates algorithm steps on the basis of K cluster centroids until the clustering results converge into K clusters. In its nature, K-Means algorithm has characteristics which make different results depending on the initial and new centroids. In this paper, we propose the modified K-Means algorithm which improves the initial and new centroids decision methodologies. By evaluating the performance of two algorithms using the 16 weighting scheme of SMART system, the modified algorithm showed $20{\%}$ better results on recall and F-measure than those of K-Means algorithm, and the document clustering results are quite improved.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.10 no.3
• /
• pp.37-48
• /
• 1986

The purpose of this study was to analyze 'the natural groupings' of subjects in order to classify highly similar somatotype for clothing construction. The sample for the study was drawn randomly out of senior high school boys in Seoul urban area. The sample size was 425 boys between age 16 and 18. Cluster analysis was more concerned with finding the hierarchical structure of subjects by three dimensional distance of stature. bust girth and sleeve length. The groups forming a partition can be subdivided into 5 and 6 sets by the hierarchical tree of the given subjects. Ward's Minimum Variance Method was applied after extraction of distance matrix by the Standardized Euclidean Distance. All of the above data was analyzed by the computer installed at Korea Advanced Institute of Science and Technology. The major findings, take for instance, of 16 age group can be summarized as follows. The results of cluster analysis of this study: 1. Cluster 1 (32 persons means $18.29\%$ of the total) is characterized with smaller bust girth than that of cluster 5, but stature and sleeve length of the cluster 1 are the largest group. 2. Cluster 2 (18 Persons means $10.29\%$ of the total) is characterized with the group of the smallest stature and sleeve length, but bust girth larger than that of cluster 3. 3. Cluster 3(35persons means $20\%$ of the total) is classified with the smallest group of all the stature, bust girth and sleeve length. 4. Cluster 4(60 persons means $34.29\%$ of the total) is grouped with the same value of sleeve length with the mean value of 16 age group, but the stature and bust girth is smaller than the mean value of this age group. 5. Cluster 5(30 persons means $17.14\%$ of the total) is characterized with smaller stature than that of cluster 1, and with larger bust girth than that of cluster 1, but with the same value of the sleeve length with the mean value of the 16 age group.

• Asia pacific journal of information systems
• /
• v.16 no.1
• /
• 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.