• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.38 no.1
• /
• pp.101-109
• /
• 2015

This paper suggests a procedure to define business process improvement (BPI) projects with analysis results based on the cause-and-effect chain. The procedure developed in this paper focuses on eliminating root causes of business problems resulted from abnormal events occurred in business process executions. First, we develop three criteria used to make clusters of the root causes where a cluster of root causes will be eliminated together by a BPI project defined based on the cluster. Second, we develop a method to formulate desired expectations from the BPI project. Also, we suggest a method to calculate the relative importance of the BPI projects that help a BPI organization determine priorities of them. We illustrate the procedure and the methods with some examples for the domestic mail delivery process in the postal service industry.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.5
• /
• pp.65-72
• /
• 2017

Premature Ventricular Contraction(PVC) arrhythmia is most common abnormal-heart rhythm that may increase mortal risk of a cardiac patient. Thus, it is very important issue to identify the specular portraits of PVC pattern especially from the patient. In this paper, we propose a new method to extract the characteristics of PVC pattern by applying K-means machine learning algorithm on Heart Rate Variability depicted in Poinecare plot. For the quantitative analysis to distinguish the trend of cluster patterns between normal sinus rhythm and PVC beat, the Euclidean distance measure was sought between the clusters. Experimental simulations on MIT-BIH arrhythmia database draw the fact that the distance measure on the cluster is valid for differentiating the pattern-traits of PVC beats. Therefore, we proposed a method that can offer the simple remedy to identify the attributes of PVC beats in terms of K-means clusters especially in the long-period Electrocardiogram(ECG).

• ETRI Journal
• /
• v.29 no.5
• /
• pp.700-702
• /
• 2007

We present a relevance feedback approach based on multi-class support vector machine (SVM) learning and cluster-merging which can significantly improve the retrieval performance in region-based image retrieval. Semantically relevant images may exhibit various visual characteristics and may be scattered in several classes in the feature space due to the semantic gap between low-level features and high-level semantics in the user's mind. To find the semantic classes through relevance feedback, the proposed method reduces the burden of completely re-clustering the classes at iterations and classifies multiple classes. Experimental results show that the proposed method is more effective and efficient than the two-class SVM and multi-class relevance feedback methods.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.30A no.3
• /
• pp.139-145
• /
• 1993

Partitioning is an important step in the hierarchical design of very large scale integrated circuits. In this research, a new effective partitioning algorithm based on 2-level hierarchy is presented. At the beginning, clusters are formed to reduce the problem size. To overcome the weakness of the iterative improvement techniques that the partitioning result is dependent on the initial partitioning and to consistently produce good results, the cluster-level partitioning is performed several times using several sets of parameters. Then the best result of cluster-partitioning is used as the initial solution for lower level partitioning. For each partitioning, the gradual constraint enforcing partitioning method has been used. The clustering-based partitioning algorithm has been applied to several benchmark examples and produced promising results which show that this algorithm is efficient and effective.

• Journal of Communications and Networks
• /
• v.14 no.5
• /
• pp.518-523
• /
• 2012

This paper presents new coherent and non-coherent detection methods for the IEEE 802.15.4a low-rate ultra-wideband physical layer with forward error correction (FEC) coding techniques. The coherent detection method involving channel estimation is based on the correlation characteristics of the preamble signal. A coherent receiver uses novel iterated selective-rake (IT-SRAKE) to detect 2-bit data in a non-line-of-sight channel. The non-coherent detection method that does not involve channel estimation employs a 2-bit data detection scheme using modified transmitted reference pulse cluster (M-TRPC) methods. To compare the two schemes, we have designed an IT-SRAKE receiver and a MTRPC receiver using an IEEE 802.15.4a physical layer. Simulation results show the performance of IT-SRAKE is better than that of the M-TRPC by 3-9 dB.

• Electrical & Electronic Materials
• /
• v.8 no.6
• /
• pp.774-779
• /
• 1995

Process optimization experiments based on the Taguchi method were performed in order to set up the optimal process conditions for the contact oxide etching process module which was built in order to be attached to the cluster system of multi-processing purpose. In order to compare with Taguchi method, the contact oxide etching process carried out with different process parameters(CHF$_{3}$/CF$_{4}$ gas flow rate, chamber pressure, RF power and magnetic field intensity). Optimal etching characteristics were evaluated in terms of etch rate, selectivity, uniformity and etched profile. In this paper, as a final analysis of experimental results the optimal etching characteristics were obtained at the process conditions of CHF3/CF4 gas flow rate = 72/8 sccm, chamber pressure = 50 mTorr, RF power = 500 watts, and magnetic field intensity = 90 gauss.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.21 no.4
• /
• pp.727-739
• /
• 1997

The purpose of this study was to classify and analyze the lower body of adult female and to provide it'stotal data for clothing construction. The subjects were 82 Korean adult females and their age range is from 19 to 24 years old. 42 variables (10 variables from the direct anthropometric data, 2 variables from the multiplication method and 30 variables from the indirect anthropometric data) in total were applies to analyze(means, standard deviations, factor analysis, cluster analysis) The result of factor analysis indicated that 10 factors were extract'cd through factor analysis and orthogonal rotation by the method of varimax and those factors comprise 82. 5 percent of total variance. The obesity of lower body was closely related to hip angle indicated the degree of drooping hip was extracted a independent factor, not influenced by other variance. And somatotype of lower body is classified by cluster analysis, using the FASTCLUS of SAS. To classify the lower body, two kinds of silhouette, front- back and side were applied to analize. The front- back silhouette was subdivided into five groups and the side silhouette four.

• Bulletin of the Korean Chemical Society
• /
• v.12 no.5
• /
• pp.574-582
• /
• 1991

The CO molecules adsorbed on Ni(111) surface is studied in the cluster approximation employing EH method with self-consistent charge iteration. The effect of CO coverage is simulated by allowing the variation of valence state ionization potentials of each Ni atom in model cluster according to the self-consistent charge iteration method. The CO coverage dependent C-O stretching frequency shift, adsorption site conversion, and metal work function change are attributed to the charge transfer between metal surface and adsorbate. For CO/Ni(111) system, net charge transfer from Ni surface to chemisorbed CO molecules makes surface Ni atoms be more positive with increasing coverage, and lowers Ni surface valence band. This leads to a weaker interaction between metal surface valence band and Co $2{\pi}^{\ast}$ MO, less charge transfer to a single CO molecule, and the bule shift of C-O stretching frequency. Further increase of coverage induces the conversion of 3-fold site CO to lower coordination site CO as well as the blue shift of C-O stretching frequency. This whole process is accompanied by the continuous increase of metal work function.

• International Journal of Computer Science & Network Security
• /
• v.22 no.9
• /
• pp.389-395
• /
• 2022

A Mobile Sensor Network is widely used in real time applications. A critical need in Mobile Sensor Network is to achieve energy efficiency during routing as the sensor nodes have scarce energy resource. The nodes' mobility in MWSN poses a challenge to design an energy efficient routing protocol. Clustering helps to achieve energy efficiency by reducing the organization complexity overhead of the network which is proportional to the number of nodes in the network. This paper proposes"EERA: Energy Efficient Routing Algorithm for Mobile Sensor Network" is divided into five phases. 1, Cluster Formation 2.Cluster head and Transmission head selection 3.Path Establishment / Route discovery and 4,Data Transmission. Experimental Analysis has been done and is found that the proposed method performs better than the existing method with respect to four parameters.

• The Bulletin of The Korean Astronomical Society
• /
• v.44 no.2
• /
• pp.74.1-74.1
• /
• 2019

We present the distance measurements of two spiral galaxies NGC 4517, NGC 4592, and neighboring dwarf galaxies found in Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP) wide field survey data. Distances to NGC 4517 and NGC 4592 are measured by the Tip of the Red Giant Branch method from archival Hubble Space Telescope data; 9.00^+0.094_-0.260 Mpc for NGC 4517 and 8.90^+0.256_-0.060 Mpc for NGC 4592. The spatial distance between NGC 4517 and NGC 4592 is 300 kpc, which is close enough for them to be considered as a group (NGC 4517 group). Using resolved stellar photometry and Surface Brightness Fluctuation (SBF) method with HSC-SSP data, we estimate the distances to three other dwarf galaxies and confirm that they are members of the group. Velocities of three of the galaxies in the NGC 4517 group show that this group is one of the galaxy groups in the near side of the Virgo Cluster infall region.