• Communications for Statistical Applications and Methods
• /
• v.24 no.5
• /
• pp.457-480
• /
• 2017

We develop a random partition procedure based on a Dirichlet process prior with Laplace distribution. Gibbs sampling of a Laplace mixture of linear mixed regressions with a Dirichlet process is implemented as a random partition model when the number of clusters is unknown. Our approach provides simultaneous partitioning and parameter estimation with the computation of classification probabilities, unlike its counterparts. A full Gibbs-sampling algorithm is developed for an efficient Markov chain Monte Carlo posterior computation. The proposed method is illustrated with simulated data and one real data of the energy efficiency of Tsanas and Xifara (Energy and Buildings, 49, 560-567, 2012).

• Journal of Sericultural and Entomological Science
• /
• v.39 no.2
• /
• pp.119-133
• /
• 1997

Isozyme was used to characterize general protein patterns of genetic relationships among 303 silkworm stocks preserved in National Sericultural and Entomology Research Institute, RDA. Six isozymes (esterase, acid phosphatase, alkaline phosphatase, amylase, glucose-6-phosphate dehydrogenase and sucrase) from hemolymph, midgut, and digestive juice were employed to construct dendograms(UPGMA method) using a polycrylamide gel electrophoresis. A cluster analysis revealed four major group, which were divided into several subgroups within each group, contained assemglages of Japanese and Chinese races. Especially, genetic differentiation in the first and second group was greatest rather than within Japanese and Chinese races repectively and was concordant with the hypothesis of phyletic sorting of initial variability in China many years ago. Hypothesized recent introgression between groups was also plausible, but the eviednce suggested bidirectional gene flow between the Chinese and the Japnaese lineages. Interpreting the results in light of evidence from the current study, the genetic diversity and relationship showed in Korean silkworm race, Hansammyun reflected early and independent evolution from the Chinese ancestor, limited addition of new variability and phyletic sorting within Korean peninsula more than 4,000 years.

• Journal of information and communication convergence engineering
• /
• v.6 no.4
• /
• pp.466-469
• /
• 2008

The existing routing protocols in USN environment, PEGASIS is more efficient than LEACH, which is a hierarchical routing protocol, for network configuration based on power consumption. Despite its merit that it can reduce energy consumption per node, however, the PEGASIS protocol also has a weakness that it is less responsive to frequent changes that occur in the configuration of sensor network due to BS nodes that keep changing, which is a typical characteristic of the sensor network. To address this problem, this paper proposes to select sub-cluster heads and have them serve as intermediate nodes. This paper presents and analyses that this method can resolve the aforementioned problem of the PEGASIS algorithm.

• Journal of Information Processing Systems
• /
• v.14 no.2
• /
• pp.377-395
• /
• 2018

Aiming at the problem of service reliability in resource reservation in cloud computing environments, a model of dynamic cloud resource reservation based on trust is proposed. A domain-specific cloud management architecture is designed in which resources are divided into different management domains according to the types of service for easier management. A dynamic resource reservation mechanism (DRRM) is used to test users' reservation requests and reserve resources for users. According to user preference, several resources are chosen to be candidate resources by fuzzy cluster analysis. The fuzzy evaluation method and a two-way trust evaluation mechanism are adopted to improve the availability and credibility of the model. An analysis and simulation experiments show that this model can increase the flexibility of resource reservation and improve user satisfaction.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2001.10a
• /
• pp.171-174
• /
• 2001

This study investigates the differences in critical dimensions which impact on operations and strategy in Internet EC of service related intangible product. For this purpose, services are newly classified by two selected dimensions such as 1)the proportion of substitute by on-line, and 2)the needs of interaction and customization. Secondly, on the classification of services, the differences of 1) customer needs of geographical accessibility, 2) needs of cooperation with off-line functions, and 3) customer purchase intention in Internet EC are tested among classified groups. Finally, implementations on operations and strategy in Internet EC are suggested, based on the results of analysis. Data are collected by the survey on the customer groups, and analyzed by statistical method, such as mean score plot, cluster analysis, and analysis of variance.

• MALSORI
• /
• no.48
• /
• pp.93-105
• /
• 2003

Since telephone channel has bandlimited frequency characteristics, speech signal over the telephone channel shows degraded speech quality. In this paper, we propose an algorithm using neural network to reconstruct wideband speech from its narrowband version. Although single neural network is a good tool for direct mapping, it has difficulty in training for vast and complicated data. To alleviate this problem, we modularize the neural networks based on appropriate clustering of the acoustic space. We also introduce fuzzy computing to compensate for probable misclassification at the cluster boundaries. According to our simulation, the proposed algorithm showed improved performance over the single neural network and conventional codebook mapping method in both objective and subjective evaluations.

• The Korean Journal of Applied Statistics
• /
• v.28 no.6
• /
• pp.1209-1216
• /
• 2015

Competing-risks events are often observed in a clustered clinical study such as a multi-center clinical trial. We propose a joint modelling approach via a shared frailty term for competing risks survival data from a cluster. For the inference we use the hierarchical likelihood (or h-likelihood), which avoids an intractable integration. We derive the corresponding h-likelihood procedure. The proposed method is illustrated via the analysis of a practical data set.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.13 no.3 s.31
• /
• pp.197-206
• /
• 1989

This study was to investigate brand images of ladies wear and the effects of fashion advertisements, and to determine the difference of brand image preference and the effects of fashion advertising by clothing life style. 421 subjects were gathered through stratified sampling method and, for data analysis, frequency distribution, $x^2-test$, Cronbach'$\alpha$, ANOVA, Duncan's multiple Range test, Multiple Dimentional Scaling (M.D.S.), Factor analysis, Cluster analysis, were conducted. The results are as follows; 1. In image formation process, feminine as formal image, manish as casual image, were recognized. 2. Four factors were determined for analysis of clothing life styles and with these factors five different clothing life style groups were classified. 3. There was a meaningful difference between clothing life style and brand image preference, and also between clothing life style and the effects of fashion advertisement. From these findings, general and specific fashion advertising strategies are proposed.

• Journal of the Korean Society of Safety
• /
• v.29 no.5
• /
• pp.136-145
• /
• 2014

This article has compared the level of activities of occupational safety and health in workplace among sub-sectors of manufacturing industry in order to set the priority for policy intervention. Data of manufacturing industry in the survey on the current status of occupational safety and health was used with factor analysis and radar graphic method. Authors have categorized sub-sectors of manufacturing industry into four categories, attained group, active group, neglected group, and passive group based on injury rate, level of safety and health activities. The neglected group may be the first target group for occupational safety and health policy guiding some detailed occupational safety and health activities. Limitation of this study is that cross sectional data was analyzed. The long term effect could not be analyzed.

• Journal of Korea Multimedia Society
• /
• v.23 no.1
• /
• pp.1-7
• /
• 2020

Clustering methods essentially take a two-step approach; extracting feature vectors for dimensionality reduction and then employing clustering algorithm on the extracted feature vectors. However, for clustering images, the traditional clustering methods such as stacked auto-encoder based k-means are not effective since they tend to ignore the local information. In this paper, we propose a method first to effectively reduce data dimensionality using convolutional auto-encoder to capture and reflect the local information and then to accurately cluster similar data samples by using a hierarchical clustering approach. The experimental results confirm that the clustering results are improved by using the proposed model in terms of clustering accuracy and normalized mutual information.