• KIISE Transactions on Computing Practices
• /
• v.21 no.6
• /
• pp.436-441
• /
• 2015

There are many types of clustering algorithms such as centroid, hierarchical, or density-based methods. Each algorithm has unique data grouping principles, which creates different varieties of clusters. Ordering Points To Identify the Clustering Structure (OPTICS) is a well-known density-based algorithm to analyze arbitrary shaped and varying density clusters, but the obtained clusters only correlate loosely. Hierarchical agglomerative clustering (HAC) reveals a hierarchical structure of clusters, but is unable to clearly find non-convex shaped clusters. In this paper, we provide a novel hierarchy generation framework and application which can aid users by combining the advantages of the two clustering methods.

• Atmosphere
• /
• v.16 no.4
• /
• pp.259-267
• /
• 2006

In the present study, the neural network (NN) model with cluster analysis method was developed to predict storm surge in the whole Korean coastal regions with special focuses on the regional extension. The model used in this study is NN model for each cluster (CL-NN) with the cluster analysis. In order to find the optimal clustering of the stations, agglomerative method among hierarchical clustering methods was used. Various stations were clustered each other according to the centroid-linkage criterion and the cluster analysis should stop when the distances between merged groups exceed any criterion. Finally the CL-NN can be constructed for predicting storm surge in the cluster regions. To validate model results, predicted sea level value from CL-NN model was compared with that of conventional harmonic analysis (HA) and of the NN model in each region. The forecast values from NN and CL-NN models show more accuracy with observed data than that of HA. Especially the statistics analysis such as RMSE and correlation coefficient shows little differences between CL-NN and NN model results. These results show that cluster analysis and CL-NN model can be applied in the regional storm surge prediction and developed forecast system.

• The Korean Journal of Applied Statistics
• /
• v.18 no.3
• /
• pp.555-565
• /
• 2005

To Know customers is very important for the company to survive in its cut-throat competition among coimpetitors. Companies need to manage the relationship with each ana every customer, ant make each of customers as profitable as possible. CRM (Customer relationship management) has emerged as a key solution for managing the profitable relationship. In order to achieve successful CRM customer segmentation is a essential component. Clustering as a data mining technique is very useful to build data-driven segmentation. This paper is concerned with building proper customer segmentation with introducing a credit card company case. Customer segmentation was built based only on transaction data which cattle from customer's activities. Two-step clustering approach which consists of k-means clustering and agglomerative clustering was applied for building a customer segmentation.

• Journal of Korean Society for Atmospheric Environment
• /
• v.13 no.5
• /
• pp.345-353
• /
• 1997

The purpose of the study was to stastically classify individual PM-10 measured by SEM/EDX (scanning electron microscopy/energy dispersive x-ray analyzer). The SEM/EDX provided various physical parameters like optical diameter, as well as major 18 chemical information (Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, Br, Pb) for a particle-by-particle basis. The total of 1,419 particles were analyzed for the study. Thus density and mass of each particle can be estimated based on its chemical composition. Further the study developed 4 semisource profiles including highway, oil boiler, incinerator, and soil emissions, where each sample was collected near the source in the ambient air The profiles developed were consisted of mass fractions and their uncertainties based on a particle class concept. To obtain mass fraction of each particle class, an agglomerative hierarchical cluster analysis was initially applied to create particle classes for each sample. Then uncertainties were calculated for each class based on the jacknife method. The 1,258 particles out of 1,419 (88.7%) were assorted in newly generated particle classes. The study provides opportunities to identify particle's source quantitatively and to develope various receptor models.

• Journal of the Korean Society for information Management
• /
• v.23 no.4 s.62
• /
• pp.215-231
• /
• 2006

Recently there are many bibliometric studies attempting to utilize Pathfinder networks(PFNets) for examining and analyzing the intellectual structure of a scholarly field. Pathfinder network scaling has many advantages over traditional multidimensional scaling, including its ability to represent local details as well as global intellectual structure. However there are some limitations in PFNets including very high time complexity. And Pathfinder network scaling cannot be combined with cluster analysis, which has been combined well with traditional multidimensional scaling method. In this paper, a new method named as Parallel Nearest Neighbor Clustering (PNNC) are proposed for complementing those weak points of PFNets. Comparing the clustering performance with traditional hierarchical agglomerative clustering methods shows that PNNC is not only a complement to PFNets but also a fast and powerful clustering method for organizing informations.

• Journal of Information Science Theory and Practice
• /
• v.8 no.2
• /
• pp.6-17
• /
• 2020

Word similarity is often measured to enhance system performance in the information retrieval field and other related areas. This paper reports on an experimental comparison of values for word similarity measures that were computed based on 50 intentionally selected words from a Reuters corpus. There were three targets, including (1) co-occurrence-based similarity measures (for which a co-occurrence frequency is counted as the number of documents or sentences), (2) context-based distributional similarity measures obtained from a latent Dirichlet allocation (LDA), nonnegative matrix factorization (NMF), and Word2Vec algorithm, and (3) similarity measures computed from the tf-idf weights of each word according to a vector space model (VSM). Here, a Pearson correlation coefficient for a pair of VSM-based similarity measures and co-occurrence-based similarity measures according to the number of documents was highest. Group-average agglomerative hierarchical clustering was also applied to similarity matrices computed by individual measures. An evaluation of the cluster sets according to an answer set revealed that VSM- and LDA-based similarity measures performed best.

• Journal of the Korean Society for Library and Information Science
• /
• v.22
• /
• pp.331-360
• /
• 1992

This research presents author co-citation analysis of the subject area in the humanities - Korean history. Three approaches to multivariate analyses were used to display the inter-author relationships in the similarity matrix. Data on co-citation of sixty seven authors for the period of 1980­1989 were extracted from the database constructed by author. The author's name, here refers to a body of writings by a person, is the unit of analysis. The data were subjected to non-metric multidimensional scaling program create two-dimensional map of authors. Authors with similarity are clustered using hierarchical agglomerative procedure and it is found that five clusters in Korean history represent primarily research specializations. Author map of Korean history reveals the first dimension corresponding to subject orientation of authors and the second dimension corresponds to research method or research style. In factor analysis, each factor reflects research specialty made up of authors, and factor locadings demonstrate the breadth or concentration of sixty seven authors' scholarly contributions on Korean history. It is demonstrated that the· specific methodology employed by this research, author co-citation analysis, is useful to represent the intellectual structure of Korean history.

• Annual Conference of KIPS
• /
• 2014.04a
• /
• pp.952-955
• /
• 2014

본 논문에서는 정보 검색 분야에서 잘 알려진 잠재 의미 분석 방법과 계층적 군집화 방법의 단점을 상호 보완하여 보다 효율적인 정보 검색을 위한 혼합형 군집화 방법을 제안한다. 먼저, 잠재 의미 분석 방법은 벡터 연산을 통하여 자동적으로 문서 내에 있는 잠재적인 의미를 찾는 정보 검색분야에서 많이 사용되는 고전적인 방법이다. 그러나 이 방법은 언어의 유의성이나 다의성으로 인하여 발생되는 백-오브-워드(bag-of-word) 문제를 가지고 있다. 두 번째 방법인 문서 군집화를 위하여 범용적으로 사용되고 있는 계층적 군집화 방법이다. 이 방법은 이를 통하여 분석된 군집의 질적 측면에서 볼 때, 여전히 단층적 군집들이 많이 형성되어 세부적인 분석을 통한 추가적인 군집화가 필요함을 알 수 있다. 따라서, 본 논문에서는 앞서 언급한 문제점을 해결하기 위하여 혼합적인 방법으로 잠재 의미 분석 방법을 이용한 응집 계층 군집화 방법을 제안한다. 제안한 방법을 이용하여 잘 알려진 두 개의 데이터에 적용하고 기존의 방법과 그 결과를 비교함으로써 군집의 질적 측면에서의 우수함을 보인다.

• Annual Conference of KIPS
• /
• 2009.04a
• /
• pp.399-400
• /
• 2009

As a part of the era of human centric applications people started to care about their well being utilizing any possible mean. This paper proposes a framework for real time on-body sensor health-care system, addresses the current issues in such systems, and utilizes an enhanced online divisive agglomerative clustering algorithm (EODAC); an algorithm that builds a top-down tree-like structure of clusters that evolves with streaming data to rationally cluster on-body sensor data and give accurate diagnoses remotely, guaranteeing high performance, and scalability. Furthermore it does not depend on the number of data points.

• Annual Conference of KIPS
• /
• 2011.04a
• /
• pp.230-233
• /
• 2011

Chronic disease is linked to patient's' lifestyle. Therefore, doctor has to monitor his/her patient over time. This may involve reviewing many reports, finding any changes, and modifying several treatments. One solution to optimize the burden is using a visualizing tool over time such as a timeline-based visualization tool where all reports and medicine are integrated in a problem centric and time-based style to enable the doctor to predict and adjust the treatment plan. This solution was proposed by Bui et. al. [2] to observe the medical history of a patient. However, there was limitation of studying the diabetes patient's history to find out what was the cause of the current development in patient's condition; moreover what would be the prediction of current implication in one of the diabetes' related factors (such as fat, cholesterol, or potassium). In this paper, we propose a Grid-based Interactive Diabetes System (GIDS) to support bioinformatics analysis application for diabetes diseases. GIDS used an agglomerative clustering algorithm as clustering correlation algorithm as primary algorithm to focus medical researcher in the findings to predict the implication of the undertaken diabetes patient. The algorithm was Chronological Clustering proposed by P. Legendre [11] [12].