• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.87-91
• /
• 2011

강수는 지역에 따라 발생양상이 매우 다른 자연현상 중 하나이다. 이러한 강수를 효과적으로 분석하여 확률강수량을 산정하기위해서 수문학에서는 다양한 방법이 시도되어 왔다. 우리나라에서는 지점빈도해석을 통한 확률강수량을 주로 사용해왔으나 최근 들어 Hosking and Wallis(1997)가 제안한 지역빈도해석을 활용을 적극 도모 하고 있는 중이다. 이러한 지역빈도해석 기법은 지점빈도해석 기법에 비하여 한정된 강수자료를 활용하는 측면 등 여러 가지 장점을 가진 확률 강수량 산정방법이다. 그러나 이 기법을 적용하여 확률강수량을 산정하기 위해서는 강수의 지역구분을 먼저 수행하여야 한다. 강수지역의 구분을 위해서는 여러 가지 기법이 존재하나 최근에는 Cluster 기법 중 K-means 방법이나 Fuzzy c-means 방법 등을 주로 적용하여 지역구분을 수행하고 있다. 그러나 K-means 방법이나 Fuzzy c-means 방법 등은 산정 방법내에서 최적 군집수를 결정할 수 있는 알고리즘이 없기 때문에 임의적으로 최적 군집수를 결정하여야 한다. 본 연구에서는 이러한 단점을 극복하기 위하여 Cluster 평가지수 중 하나인 Dunn 지수를 이용하여 최적 군집수를 제시하고자 한다. 본 연구에서 강수지역을 구분하기 위하여 적용한 인자는 월 평균 강수량, 연 평균 강수량, 월 최대 강수량, 경도, 위도, 고도 등이며, 이를 K-means, PAM 및 친근도 전파 기법을 통하여 강수지역을 구분하였다. 적정 군집수를 임의적으로 증가시켜 가면서 Dunn 지수를 산정하였다. 산정된 결과를 통하여 최적 군집수를 결정하였다.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.20 no.11
• /
• pp.959-964
• /
• 2007

To investigate the influence of partial discharge(PD) distribution characteristics due to various defects on the power cable joints interface, we used the K-means clustering method. As the result of PD number(n) distribution analyzing on $\Phi-n$ graph, the phase angle($\Phi$) of cluster centroid shifted to $0^{\circ}\;and\;180^{\circ}$ increasing with applying voltage. It was confirmed that the PD quantify(q) and euclidean distance of centroid were increased with applying voltage from the centroid distribution analyzing of $\Phi-q$ plane. The dispersion degree was increased with calculated standard deviation of the $\Phi-q$ cluster centroid. The PD number and mean value on $\Phi-q$ graph were some different by electric field concentration with defect types.

• Journal of Information Processing Systems
• /
• v.14 no.6
• /
• pp.1438-1444
• /
• 2018

Recently, with the development of Internet technologies and propagation of smart devices, use of microblogs such as Facebook, Twitter, and Instagram has been rapidly increasing. Many users check for new information on microblogs because the content on their timelines is continually updating. Therefore, clustering algorithms are necessary to arrange the content of microblogs by grouping them for a user who wants to get the newest information. However, microblogs have word limits, and it has there is not enough information to analyze for content clustering. In this paper, we propose a semantic-based K-means clustering algorithm that not only measures the similarity between the data represented as a vector space model, but also measures the semantic similarity between the data by exploiting the TagCluster for clustering. Through the experimental results on the RepLab2013 Twitter dataset, we show the effectiveness of the semantic-based K-means clustering algorithm.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.447-450
• /
• 2004

This paper proposes remotely sensed image classification method by fuzzy c-means clustering algorithm using average intra-cluster distance. The average intra-cluster distance acquires an average of the vector set belong to each cluster and proportionates to its size and density. We perform classification according to pixel's membership grade by cluster center of fuzzy c-means clustering using the mean-values of training data about each class. Fuzzy c-means algorithm considered membership degree for inter-cluster of each class. And then, we validate degree of overlap between clusters. A pixel which has a high degree of overlap applies to the maximum likelihood classification method. Finally, we decide category by comparing with fuzzy membership degree and likelihood rate. The proposed method is applied to IKONOS remote sensing satellite image for the verifying test.

• Proceedings of the Korea Contents Association Conference
• /
• 2009.05a
• /
• pp.792-796
• /
• 2009

Determining the optimal number of cluster is an important issue in research area of data clustering. It is choosing the cluster validity method and finding the cluster number where it optimizes the cluster validity. In this paper, an efficient heuristic for determining optimal number of cluster using clustering balance is proposed. The experimental results using k-means at artificial and real-life data set show that proposed algorithm is excellent in aspect of time efficiency.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2007.11a
• /
• pp.203-204
• /
• 2007

This paper investigated the influence on partial discharge distribution of various defects at the model power cable joints interface using K-means clustering. As the result of analyzing discharge number distribution of ${\Phi}-n$ cluster, clusters shifted to $0^{\circ}\;and\;180^{\circ}$ with increasing applying voltage. It was confirmed that discharge quantity and euclidean distance between centroids were increased with applying voltage from the analyzing centroid distribution of ${\Phi}-q$ cluster. The degree of dispersion was increased with calculating standard deviation of ${\Phi}-q$ cluster centroid. The tendency both number of discharge and mean value of ${\Phi}-q$ cluster centroid were some different with defect types.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.3
• /
• pp.354-359
• /
• 2008

Clustering is a method to classify the given data set with same property into several classes. To cluster data, many methods such as K-Means, Fuzzy C-Means(FCM), Mountain Method(MM), and etc, have been proposed and used. But the clustering results of conventional methods are sensitively influenced by initial values given for clustering in each method. Especially, FCM is very sensitive to noisy data, and cluster center distortion phenomenon is occurred because the method dose clustering through minimization of within-clusters variance. In this paper, we propose a clustering method which reduces cluster center distortion through merging the nearest data based on the data weight, and not being influenced by initial values. We show the effectiveness of the proposed through experimental results applied it to various types of data sets, and comparison of cluster centers with those of FCM.

• Journal of the Korea Convergence Society
• /
• v.12 no.7
• /
• pp.169-179
• /
• 2021

This study shows that the two-stage k-means clustering method can improve prediction performance by predicting the stock price, To this end, this study introduces the two-stage k-means clustering algorithm and tests the prediction performance through comparison with various machine learning techniques. It selects the cluster close to the prediction target obtained from the k-means clustering, and reapplies the k-means clustering method to the cluster to search for a cluster closer to the actual value. As a result, the predicted value of this method is shown to be closer to the actual stock price than the predicted values of other machine learning techniques. Furthermore, it shows a relatively stable predicted value despite the use of a relatively small cluster. Accordingly, this method can simultaneously improve the accuracy and stability of prediction, and it can be considered as the new clustering method useful for small data. In the future, developing the two-stage k-means clustering is required for the large-scale data application.

• Journal of Digital Contents Society
• /
• v.12 no.3
• /
• pp.339-345
• /
• 2011

The k-means has been one of the popular, simple and faster clustering algorithms, but the right value of k is unknown. The value of k (the number of clusters) is a very important element because the result of clustering is different depending on it. In this paper, we present a novel algorithm based on an external cluster relationship criterion which is an evaluation metric of clustering result to determine the number of clusters dynamically. Experimental results show that our algorithm is superior to other methods in terms of the accuracy of the number of clusters.

• Journal of Internet Computing and Services
• /
• v.12 no.2
• /
• pp.103-111
• /
• 2011

Clustering methods is divided into hierarchical clustering, partitioning clustering, and more. If the amount of documents is huge, it takes too much time to cluster them in hierarchical clustering. In this paper we deal with K-Means algorithm that is one of partitioning clustering and is adequate to cluster so many documents rapidly and easily. We propose the new method of selecting initial seeds in K-Means algorithm. In this method, the initial seeds have been selected that are positioned as far away from each other as possible.