• Journal of KIISE:Software and Applications
• /
• v.35 no.9
• /
• pp.572-587
• /
• 2008

The Self-organizing Feature Map(SOFM) that is one of unsupervised neural networks is a very powerful tool for data clustering and visualization in high-dimensional data sets. Although the SOFM has been applied in many engineering problems, it needs to cluster similar weights into one class on the trained SOFM as a post-processing, which is manually performed in many cases. The traditional clustering algorithms, such as t-means, on the trained SOFM however do not yield satisfactory results, especially when clusters have arbitrary shapes. This paper proposes automatic clustering on trained SOFM, which can deal with arbitrary cluster shapes and be globally optimized by graph cuts. When using the graph cuts, the graph must have two additional vertices, called terminals, and weights between the terminals and vertices of the graph are generally set based on data manually obtained by users. The Proposed method automatically sets the weights based on mode-seeking on a distance matrix. Experimental results demonstrated the effectiveness of the proposed method in texture segmentation. In the experimental results, the proposed method improved precision rates compared with previous traditional clustering algorithm, as the method can deal with arbitrary cluster shapes based on the graph-theoretic clustering.

• Journal of Korea Multimedia Society
• /
• v.14 no.11
• /
• pp.1438-1449
• /
• 2011

UWSNs(Underwater Wireless Sensor Networks) need effective modeling fitted to the customized type of application and its covering area. In particular it requires an energy efficient data aggregation method for such customized application. In this paper, we envisage the application oriented model for monitoring the pollution or intrusion detection over a given underwater area. The suggested model is based on the honeycomb array of hexagonal prisms. In this model, the purpose of data aggregation is that the head node of each layer(cluster) receives just one event data arrived firstly and transfer this and its position data to the base station effectively in the manner of energy efficiency and simplicity without duplication. Here if we apply the existent data aggregation methods to this kind of application, the result is far from energy efficiency due to the complexity of the data aggregation process based on the shortest path or multicast tree. In this paper we propose three energy efficient and simple data aggregation methods in the domain of cluster and three in the domain of inter-cluster respectively. Based on the comparative performance analysis of the possible combination pairs in the two domains, we derive the best energy efficient data aggregation method for the suggested application.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.02a
• /
• pp.96-97
• /
• 2012

In nitride and oxide film deposition, sputtered metals react with nitrogen or oxygen gas in a vacuum chamber to form metal nitride or oxide films on a substrate. The physical properties of sputtered films (metals, oxides, and nitrides) are strongly influenced by magnetron plasma density during the deposition process. Typical target power densities on the magnetron during the deposition process are ~ (5-30) W/cm2, which gives a relatively low plasma density. The main challenge in reactive sputtering is the ability to generate a stable, arc free discharge at high plasma densities. Arcs occur due to formation of an insulating layer on the target surface caused by the re-deposition effect. One current method of generating an arc free discharge is to use the commercially available Pinnacle Plus+ Pulsed DC plasma generator manufactured by Advanced Energy Inc. This plasma generator uses a positive voltage pulse between negative pulses to attract electrons and discharge the target surface, thus preventing arc formation. However, this method can only generate low density plasma and therefore cannot allow full control of film properties. Also, after long runs ~ (1-3) hours, depends on duty cycle the stability of the reactive process is reduced due to increased probability of arc formation. Between 1995 and 1999, a new way of magnetron sputtering called HIPIMS (highly ionized pulse impulse magnetron sputtering) was developed. The main idea of this approach is to apply short ${\sim}(50-100){\mu}s$ high power pulses with a target power densities during the pulse between ~ (1-3) kW/cm2. These high power pulses generate high-density magnetron plasma that can significantly improve and control film properties. From the beginning, HIPIMS method has been applied to reactive sputtering processes for deposition of conductive and nonconductive films. However, commercially available HIPIMS plasma generators have not been able to create a stable, arc-free discharge in most reactive magnetron sputtering processes. HIPIMS plasma generators have been successfully used in reactive sputtering of nitrides for hard coating applications and for Al2O3 films. But until now there has been no HIPIMS data presented on reactive sputtering in cluster tools for semiconductors and MEMs applications. In this presentation, a new method of generating an arc free discharge for reactive HIPIMS using the new Cyprium plasma generator from Zpulser LLC will be introduced. Data (or evidence) will be presented showing that arc formation in reactive HIPIMS can be controlled without applying a positive voltage pulse between high power pulses. Arc-free reactive HIPIMS processes for sputtering AlN, TiO2, TiN and Si3N4 on the Applied Materials ENDURA 200 mm cluster tool will be presented. A direct comparison of the properties of films sputtered with the Advanced Energy Pinnacle Plus + plasma generator and the Zpulser Cyprium plasma generator will be presented.

• Journal of Preventive Medicine and Public Health
• /
• v.31 no.1 s.60
• /
• pp.59-71
• /
• 1998

To examine the association of hypertension with cluster of obesity, abnormal glucose and dyslipidemia in Korean urban population, we conducted this cross-sectional study among 3027 men and 2127 women age 20-85 years who visited a prevention center between May 1991 and June 1995 for a multiphasic health check at St. Mary's Hospital, Seoul. By the self-administered questionnaire, the informations of educational attainments, monthly income, alcohol consumption, cigarette smoking, and physical excercise level were obtained. Height, weight, and blood pressure were measured by a trained nurse. The fasting blood sugar (FBS), total cholesterol, high density lipoprotein (HDL) cholesterol and triglyceride were tested by enzyme method. Low density lipoprotein (LDL) cholesterol was calculated by 'total cholesterol - HDL cholesterol - triglyceride/5'. For testing the differences of cardiovascular risk factors between hypertension and normotension group, 1-test and $\chi^2$-test were performed. For the age adjusted odds ratios of hypertension in persons with obesity, abnormal glucose, and dyslipidemia compared with normal, logistic regression was performed by using SAS pakageprograme. The results obtained were as follows: 1. Age, weight, body mass index(BMI), blood glucose, total cholesterol, LDL cholesterol, and triglyceride of hypertension group in men and women were significantly higher than normotension group, but height and HDL cholesterol of hypertension group only in women significantly lower than normotension group. The frequency of obesity $(BMI\geq25kg/m^2)$, abnormal glucose $(\geq\;120mg/dl)$, hypercholesterolemia $(\geq\;240mg/dl)$, lower HDL cholesterol (<45 mg/dl in women only), higher LDL cholesterol $(\geq\;160mg/dl)$, and hyper hypertriglyceridemia $(\geq\;250mg/dl)$ in hypertension group of men and women were significantly higher than normotension group. 2. Systolic and diastolic blood pressure were negatively correlated with hight, but positively with age, weight, BMI, total cholesterol, LDL cholesterol, and triglyceride in men and women. BMI was positively correlated with fasting blood sugar, total cholesterol, LDL cholesterol and triglyceride but negatively with HDL cholesterol. 3. The age adjusted odds ratios of hypertension were as follows in men and women : among persons who were obese compared with those nonobese, 2.53 (95% Confidence Intervals [C.I.] 2.08-3.07) and 2.22 (95%C.I. 1.71-2.87): among persons who were abnormal glucose compared with those normoglycemic, 1.43 (95%C.I 1.13-1.82) and 2.01 (95%C.I 1.36-2.94): and among persons who were dyslipidemia (hypercholesterolemia or lower HDL cholesterol or higher LDL cholesterol or hypertriglyceridemia) compared with those normal lipid, 1.59 (95%C.I 1.30-1.95) and 1.51 (95%C.I 1.16-1.96). After combined more than one risk factor, the odds ratios were increased. Among persons with cluster of obesity, abnormal glucose, and dyslipidemia, the odds ratio of hypertension was 2.25 (95%C.I 1.47-3.37) in men and 3.02 (95%C.I 1.71-5.30) in women. In conclusion, it was suggested that hypertension was associated with cluster of obesity, abnormal glucose, dyslipidemia in this Korean urban population.

• Journal of the Korea Computer Graphics Society
• /
• v.16 no.4
• /
• pp.25-33
• /
• 2010

With the popularization of digital photography, people are now capturing and storing far more photos than ever before. However, the enormous number of photos often discourages the users to identify desired photos. In this paper, we present a novel method for fast and intuitive browsing through large collections of geo-referenced photographs. Given a set of photos, we construct a hierarchical structure of clusters such that each cluster includes a set of spatially adjacent photos and its sub-clusters divide the photo set disjointly. For each cluster, we pre-compute its convex hull and the corresponding polygon area. At run-time, this pre-computed data allows us to efficiently visualize only a fraction of the clusters that are inside the current view and have easily recognizable sizes with respect to the current zoom level. Each cluster is displayed as a single polygon representing its convex hull instead of every photo location included in the cluster. The users can quickly transfer from clusters to clusters by simply selecting any interesting clusters. Our system automatically pans and zooms the view until the currently selected cluster fits precisely into the view with a moderate size. Our user study demonstrates that these new visualization and interaction techniques can significantly improve the capability of navigating over large collections of geo-referenced photos.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.8 no.6
• /
• pp.460-469
• /
• 2015

This study's aim is for reducing big data transfer time of IoT healthcare devices by modulating digital bits into Manchester code including zero-voltage idle as information for secondary compressed binary cluster's compartment after two step compression of compressing binary data into primary and secondary binary compressed clusters for each binary clusters having compression benefit of 1 bit or 2 bits. Also this study proposed that as department information of compressed binary clusters, inserting idle signal into Manchester code will have benefit of reducing transfer time in case of compressing binary cluster into secondary compressed binary cluster by 2 bits, because in spite of cost of 1 clock idle, another 1 bit benefit can play a role of reducing 1 clock transfer time. Idle signal is also never consecutive because the signal is for compartment information between two adjacent secondary compressed binary cluster. Voltage transition on basic rule of Manchester code is remaining while inserting idle signal, so DC balance can be guaranteed. This study's simulation result said that even compressed binary data by another compression algorithms could be transferred faster by as much as about 12.6 percents if using this method.

• Journal of Korean Society of Water and Wastewater
• /
• v.37 no.5
• /
• pp.261-269
• /
• 2023

In light of recent social concerns related to issues such as water supply pipe deterioration leading to problems like leaks and degraded water quality, the significance of maintenance efforts to enhance water source quality and ensure a stable water supply has grown substantially. In this study, scan statistic was applied to analyze water quality complaints and water leakage accidents from 2015 to 2021 to present a reasonable method to identify areas requiring improvement in water management. SaTScan, a spatio-temporal statistical analysis program, and ArcGIS were used for spatial information analysis, and clusters with high relative risk (RR) were determined using the maximum log-likelihood ratio, relative risk, and Monte Carlo hypothesis test for I city, the target area. Specifically, in the case of water quality complaints, the analysis results were compared by distinguishing cases occurring before and after the onset of "red water." The period between 2015 and 2019 revealed that preceding the occurrence of red water, the leak cluster at location L2 posed a significantly higher risk (RR: 2.45) than other regions. As for water quality complaints, cluster C2 exhibited a notably elevated RR (RR: 2.21) and appeared concentrated in areas D and S, respectively. On the other hand, post-red water incidents of water quality complaints were predominantly concentrated in area S. The analysis found that the locations of complaint clusters were similar to those of red water incidents. Of these, cluster C7 exhibited a substantial RR of 4.58, signifying more than a twofold increase compared to pre-incident levels. A kernel density map analysis was performed using GIS to identify priority areas for waterworks management based on the central location of clusters and complaint cluster RR data.

• International Journal of Computer Science & Network Security
• /
• v.24 no.9
• /
• pp.30-40
• /
• 2024

Building an accurate 3-D spatial road network model has become an active area of research now-a-days that profess to be a new paradigm in developing Smart roads and intelligent transportation system (ITS) which will help the public and private road impresario for better road mobility and eco-routing so that better road traffic, less carbon emission and road safety may be ensured. Dealing with such a large scale 3-D road network data poses challenges in getting accurate elevation information of a road network to better estimate the CO₂ emission and accurate routing for the vehicles in Internet of Vehicle (IoV) scenario. Clustering and regression techniques are found suitable in discovering the missing elevation information in 3-D spatial road network dataset for some points in the road network which is envisaged of helping the public a better eco-routing experience. Further, recently Explainable Artificial Intelligence (xAI) draws attention of the researchers to better interprete, transparent and comprehensible, thus enabling to design efficient choice based models choices depending upon users requirements. The 3-D road network dataset, comprising of spatial attributes (longitude, latitude, altitude) of North Jutland, Denmark, collected from publicly available UCI repositories is preprocessed through feature engineering and scaling to ensure optimal accuracy for clustering and regression tasks. K-Means clustering and regression using Support Vector Machine (SVM) with radial basis function (RBF) kernel are employed for 3-D road network analysis. Silhouette scores and number of clusters are chosen for measuring cluster quality whereas error metric such as MAE ( Mean Absolute Error) and RMSE (Root Mean Square Error) are considered for evaluating the regression method. To have better interpretability of the Clustering and regression models, SHAP (Shapley Additive Explanations), a powerful xAI technique is employed in this research. From extensive experiments , it is observed that SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions SHAP analysis validated the importance of latitude and altitude in predicting longitude, particularly in the four-cluster setup, providing critical insights into model behavior and feature contributions with an accuracy of 97.22% and strong performance metrics across all classes having MAE of 0.0346, and MSE of 0.0018. On the other hand, the ten-cluster setup, while faster in SHAP analysis, presented challenges in interpretability due to increased clustering complexity. Hence, K-Means clustering with K=4 and SVM hybrid models demonstrated superior performance and interpretability, highlighting the importance of careful cluster selection to balance model complexity and predictive accuracy.

• Proceedings of the KIEE Conference
• /
• 1991.07a
• /
• pp.426-431
• /
• 1991

Diakoptic method applied to analyze large power system, always require the efficient tearing algorithm. But conventional tearing methods is not suitable to apply practical power system. This paper presents new tearing algorithm based on factorization path concept of sparse vector method, and applied MPRLD, a kind of optimal ordering algorithm, in ordering step to improve the efficiency of tearing algorithm. Test result of model systems shows that new proposed method in this paper is enable to tear power systems not to be teared by heuristic cluster method, reduces computing time and memory size.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2003.05a
• /
• pp.729-733
• /
• 2003

The presence of local features within clusters incurred by multi-modal nature of data prohibits many conventional clustering techniques from working properly. Especially, the clustering of datasets with non-Gaussian distributions within a cluster can be problematic when the technique with implicit assumption of Gaussian distribution is used. Current study proposes a simple tandem clustering method composed of k-means type algorithm and hierarchical method to solve such problems. The multi-modal dataset is first divided into many small pre-clusters by k-means or fuzzy k-means algorithm. The pre-clusters found from the first step are to be clustered again using agglomerative hierarchical clustering method with Kullback- Leibler divergence as the measure of dissimilarity. This method is not only effective at extracting the multi-modal clusters but also fast and easy in terms of computation complexity and relatively robust at the presence of outliers. The performance of the proposed method was evaluated on three generated datasets and six sets of publicly known real world data.