• Journal of Intelligence and Information Systems
• /
• v.23 no.2
• /
• pp.19-38
• /
• 2017

Collaborative filtering (CF) algorithm has been popularly used for implementing recommender systems. Until now, there have been many prior studies to improve the accuracy of CF. Among them, some recent studies adopt 'hybrid recommendation approach', which enhances the performance of conventional CF by using additional information. In this research, we propose a new hybrid recommender system which fuses CF and the results from the social network analysis on trust and distrust relationship networks among users to enhance prediction accuracy. The proposed algorithm of our study is based on memory-based CF. But, when calculating the similarity between users in CF, our proposed algorithm considers not only the correlation of the users' numeric rating patterns, but also the users' in-degree centrality values derived from trust and distrust relationship networks. In specific, it is designed to amplify the similarity between a target user and his or her neighbor when the neighbor has higher in-degree centrality in the trust relationship network. Also, it attenuates the similarity between a target user and his or her neighbor when the neighbor has higher in-degree centrality in the distrust relationship network. Our proposed algorithm considers four (4) types of user relationships - direct trust, indirect trust, direct distrust, and indirect distrust - in total. And, it uses four adjusting coefficients, which adjusts the level of amplification / attenuation for in-degree centrality values derived from direct / indirect trust and distrust relationship networks. To determine optimal adjusting coefficients, genetic algorithms (GA) has been adopted. Under this background, we named our proposed algorithm as SNACF-GA (Social Network Analysis - based CF using GA). To validate the performance of the SNACF-GA, we used a real-world data set which is called 'Extended Epinions dataset' provided by 'trustlet.org'. It is the data set contains user responses (rating scores and reviews) after purchasing specific items (e.g. car, movie, music, book) as well as trust / distrust relationship information indicating whom to trust or distrust between users. The experimental system was basically developed using Microsoft Visual Basic for Applications (VBA), but we also used UCINET 6 for calculating the in-degree centrality of trust / distrust relationship networks. In addition, we used Palisade Software's Evolver, which is a commercial software implements genetic algorithm. To examine the effectiveness of our proposed system more precisely, we adopted two comparison models. The first comparison model is conventional CF. It only uses users' explicit numeric ratings when calculating the similarities between users. That is, it does not consider trust / distrust relationship between users at all. The second comparison model is SNACF (Social Network Analysis - based CF). SNACF differs from the proposed algorithm SNACF-GA in that it considers only direct trust / distrust relationships. It also does not use GA optimization. The performances of the proposed algorithm and comparison models were evaluated by using average MAE (mean absolute error). Experimental result showed that the optimal adjusting coefficients for direct trust, indirect trust, direct distrust, indirect distrust were 0, 1.4287, 1.5, 0.4615 each. This implies that distrust relationships between users are more important than trust ones in recommender systems. From the perspective of recommendation accuracy, SNACF-GA (Avg. MAE = 0.111943), the proposed algorithm which reflects both direct and indirect trust / distrust relationships information, was found to greatly outperform a conventional CF (Avg. MAE = 0.112638). Also, the algorithm showed better recommendation accuracy than the SNACF (Avg. MAE = 0.112209). To confirm whether these differences are statistically significant or not, we applied paired samples t-test. The results from the paired samples t-test presented that the difference between SNACF-GA and conventional CF was statistical significant at the 1% significance level, and the difference between SNACF-GA and SNACF was statistical significant at the 5%. Our study found that the trust/distrust relationship can be important information for improving performance of recommendation algorithms. Especially, distrust relationship information was found to have a greater impact on the performance improvement of CF. This implies that we need to have more attention on distrust (negative) relationships rather than trust (positive) ones when tracking and managing social relationships between users.

• Journal of Chest Surgery
• /
• v.42 no.5
• /
• pp.566-575
• /
• 2009

Background: The function of a bioprosthetic heart valve is determined largely by the material properties of the valve cusps. The uniaxial tensile test has been studied extensively. This type of testing, however, does not replicate the natural biaxial loading condition. The objective of the present study was to investigate the regional variability of the biaxial strain versus pressure relationship based on the types of fixation liquid models. Material and Method: Porcine aortic valves and pulmonary valves were assigned to three groups: the untreated fresh group, the fixed with glutaraldehyde (GA) group, and the glutaraldehyde with solvent (e.g., ethanol) group. For each group we measured the radial and circumferential stretch characteristics of the valve as a function of pressure change. Result: Radial direction elasticity of porcine aortic and pulmonary valves were better than circumferential direction elasticity in fresh, GA fixed and GA+solvent fixed groups (p=0.00). Radial and circumferential direction elasticity of pulmonary valves were better than aortic valves in GA fixed, and GA+solvent fixed groups (p=0.00). Radial and circumferential direction elasticity of aortic valves were decreased after GA and GA+solvent fixation(p=0.00), except for circumferential elasticity of GA+solvent fixed valves (p=0.785). The radial (p=0.137) and circumferential (p=0.785) direction of elasticity of aortic valves were not significantly different between GA fixed. and GA+solvent fixed groups. Radial (p=0.910) and circumferential (p=0.718) direction of elasticity of pulmonary valve also showed no significant difference between GA fixed and GA+solvent fixed groups. Conclusion: When fixing porcine valves with GA, adding a solvent does not cause a loss of mechanical properties, but, does not improve elasticity either. Radial direction elasticity of porcine aortic and pulmonary valves was better than circumferential direction elasticity.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.18 no.4
• /
• pp.221-232
• /
• 2016

Models to estimate solar radiation have been used because solar radiation is measured at a smaller number of weather stations than other variables including temperature and rainfall. For example, solar radiation has been estimated using the Angstrom-Prescott (AP) model that depends on two coefficients obtained empirically at a specific site ($AP_{Choi}$) or for a climate zone ($AP_{Frere}$). The objective of this study was to identify the coefficients of the AP model for reliable estimation of solar radiation under a wide range of spatial and temporal conditions. A global optimization was performed for a range of AP coefficients to identify the values of $AP_{max}$ that resulted in the greatest degree of agreement at each of 20 sites for a given month during 30 years. The degree of agreement was assessed using the value of Concordance Correlation Coefficient (CCC). When $AP_{Frere}$ was used to estimate solar radiation, the values of CCC were relatively high for conditions under which crop growth simulation would be performed, e.g., at rural sites during summer. The statistics for $AP_{Frere}$ were greater than those for $AP_{Choi}$ although $AP_{Frere}$ had the smaller statistics than $AP_{max}$ did. The variation of CCC values was small over a wide range of AP coefficients when those statistics were summarized by site. $AP_{Frere}$ was included in each range of AP coefficients that resulted in reasonable accuracy of solar radiation estimates by site, year, and month. These results suggested that $AP_{Frere}$ would be useful to provide estimates of solar radiation as an input to crop models in Korea. Further studies would be merited to examine feasibility of using $AP_{Frere}$ to obtain gridded estimates of solar radiation at a high spatial resolution under a complex terrain in Korea.

• Journal of Intelligence and Information Systems
• /
• v.18 no.2
• /
• pp.29-45
• /
• 2012

Bond rating is regarded as an important event for measuring financial risk of companies and for determining the investment returns of investors. As a result, it has been a popular research topic for researchers to predict companies' credit ratings by applying statistical and machine learning techniques. The statistical techniques, including multiple regression, multiple discriminant analysis (MDA), logistic models (LOGIT), and probit analysis, have been traditionally used in bond rating. However, one major drawback is that it should be based on strict assumptions. Such strict assumptions include linearity, normality, independence among predictor variables and pre-existing functional forms relating the criterion variablesand the predictor variables. Those strict assumptions of traditional statistics have limited their application to the real world. Machine learning techniques also used in bond rating prediction models include decision trees (DT), neural networks (NN), and Support Vector Machine (SVM). Especially, SVM is recognized as a new and promising classification and regression analysis method. SVM learns a separating hyperplane that can maximize the margin between two categories. SVM is simple enough to be analyzed mathematical, and leads to high performance in practical applications. SVM implements the structuralrisk minimization principle and searches to minimize an upper bound of the generalization error. In addition, the solution of SVM may be a global optimum and thus, overfitting is unlikely to occur with SVM. In addition, SVM does not require too many data sample for training since it builds prediction models by only using some representative sample near the boundaries called support vectors. A number of experimental researches have indicated that SVM has been successfully applied in a variety of pattern recognition fields. However, there are three major drawbacks that can be potential causes for degrading SVM's performance. First, SVM is originally proposed for solving binary-class classification problems. Methods for combining SVMs for multi-class classification such as One-Against-One, One-Against-All have been proposed, but they do not improve the performance in multi-class classification problem as much as SVM for binary-class classification. Second, approximation algorithms (e.g. decomposition methods, sequential minimal optimization algorithm) could be used for effective multi-class computation to reduce computation time, but it could deteriorate classification performance. Third, the difficulty in multi-class prediction problems is in data imbalance problem that can occur when the number of instances in one class greatly outnumbers the number of instances in the other class. Such data sets often cause a default classifier to be built due to skewed boundary and thus the reduction in the classification accuracy of such a classifier. SVM ensemble learning is one of machine learning methods to cope with the above drawbacks. Ensemble learning is a method for improving the performance of classification and prediction algorithms. AdaBoost is one of the widely used ensemble learning techniques. It constructs a composite classifier by sequentially training classifiers while increasing weight on the misclassified observations through iterations. The observations that are incorrectly predicted by previous classifiers are chosen more often than examples that are correctly predicted. Thus Boosting attempts to produce new classifiers that are better able to predict examples for which the current ensemble's performance is poor. In this way, it can reinforce the training of the misclassified observations of the minority class. This paper proposes a multiclass Geometric Mean-based Boosting (MGM-Boost) to resolve multiclass prediction problem. Since MGM-Boost introduces the notion of geometric mean into AdaBoost, it can perform learning process considering the geometric mean-based accuracy and errors of multiclass. This study applies MGM-Boost to the real-world bond rating case for Korean companies to examine the feasibility of MGM-Boost. 10-fold cross validations for threetimes with different random seeds are performed in order to ensure that the comparison among three different classifiers does not happen by chance. For each of 10-fold cross validation, the entire data set is first partitioned into tenequal-sized sets, and then each set is in turn used as the test set while the classifier trains on the other nine sets. That is, cross-validated folds have been tested independently of each algorithm. Through these steps, we have obtained the results for classifiers on each of the 30 experiments. In the comparison of arithmetic mean-based prediction accuracy between individual classifiers, MGM-Boost (52.95%) shows higher prediction accuracy than both AdaBoost (51.69%) and SVM (49.47%). MGM-Boost (28.12%) also shows the higher prediction accuracy than AdaBoost (24.65%) and SVM (15.42%)in terms of geometric mean-based prediction accuracy. T-test is used to examine whether the performance of each classifiers for 30 folds is significantly different. The results indicate that performance of MGM-Boost is significantly different from AdaBoost and SVM classifiers at 1% level. These results mean that MGM-Boost can provide robust and stable solutions to multi-classproblems such as bond rating.

• Korean Journal of Hospice Care
• /
• v.1 no.1
• /
• pp.46-69
• /
• 2001

Hospice Care is the best way to care for terminally ill patients and their family members. However most of them can not receive the appropriate hospice service because the Korean health delivery system is mainly be focussed on acutly ill patients. This study was carried out to clarify the situation of hospice in Korea and to develop a hospice care delivery system model which is appropriate in the Korean context. The theoretical framework of this study that hospice care delivery system is composed of hospice resources with personnel, facilities, etc., government and non-government hospice organization, hospice finances, hospice management and hospice delivery, was taken from the Health Delivery System of WHO(1984). Data was obtained through data analysis of litreature, interview, questionairs, visiting and Delphi Technique, from October 1998 to April 1999 involving 56 hospices, 1 hospice research center, 3 non-government hospice organizations, 20 experts who have had hospice experience for more than 3 years(mean is 9 years and 5 months) and officials or members of 3 non-government hospice organizations. There are 61 hospices in Korea. Even though hospice personnel have tried to study and to provide qualified hospice serices, there is nor any formal hospice linkage or network in Korea. This is the result of this survey made to clarify the situation of Korean hospice. Results of the study by Delphi Technique were as follows: 1.Hospice Resources: Key hospice personnel were found to be hospice coordinator, doctor, nurse, clergy, social worker, volunteers. Necessary qualifications for all personnel was that they conditions were resulted as have good health, receive hospice education and have communication skills. Education for hospice personnel is divided into (i)basic training and (ii)special education, e.g. palliative medicine course for hospice specialist or palliative care course in master degree for hospice nurse specialist. Hospice facilities could be developed by adding a living room, a space for family members, a prayer room, a church, an interview room, a kitchen, a dining room, a bath facility, a hall for music, art or work therapy, volunteers' room, garden, etc. to hospital facilities. 2.Hospice Organization: Whilst there are three non-government hospice organizations active at present, in the near future an hospice officer in the Health&Welfare Ministry plus a government Hospice body are necessary. However a non-government council to further integrate hospice development is also strongly recommended. 3.Hospice Finances: A New insurance standards, I.e. the charge for hospice care services, public information and tax reduction for donations were found suggested as methods to rise the hospice budget. 4.Hospice Management: Two divisions of hospice management/care were considered to be necessary in future. The role of the hospice officer in the Health & Welfare Ministry would be quality control of hospice teams and facilities involved/associated with hospice insurance standards. New non-government integrating councils role supporting the development of hospice care, not insurance covered. 5.Hospice delivery: Linkage&networking between hospice facilities and first, second, third level medical institutions are needed in order to provide varied and continous hospice care. Hospice Acts need to be established within the limits of medical law with regards to standards for professional staff members, educational programs, etc. The results of this study could be utilizes towards the development to two hospice care delivery system models, A and B. Model A is based on the hospital, especially the hospice unit, because in this setting is more easily available the new medical insurance for hospice care. Therefore a hospice team is organized in the hospital and may operate in the hospice unit and in the home hospice care service. After Model A is set up and operating, Model B will be the next stage, in which medical insurance cover will be extended to home hospice care service. This model(B) is also based on the hospital, but the focus of the hospital hospice unit will be moved to home hospice care which is connected by local physicians, national public health centers, community parties as like churches or volunteer groups. Model B will contribute to the care of terminally ill patients and their family members and also assist hospital administrators in cost-effectiveness.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.19 no.3
• /
• pp.153-163
• /
• 2017

The objective of this study is to enhance the model's speed of estimating weather variables (e.g., minimum/maximum temperature, sunshine hour, PRISM (Parameter-elevation Regression on Independent Slopes Model) based precipitation), which are applied to the Agrometeorological Early Warning System (http://www.agmet.kr). The current process of weather estimation is operated on high-performance multi-core CPUs that have 8 physical cores and 16 logical threads. Nonetheless, the server is not even dedicated to the handling of a single county, indicating that very high overhead is involved in calculating the 10 counties of the Seomjin River Basin. In order to reduce such overhead, several cache and parallelization techniques were used to measure the performance and to check the applicability. Results are as follows: (1) for simple calculations such as Growing Degree Days accumulation, the time required for Input and Output (I/O) is significantly greater than that for calculation, suggesting the need of a technique which reduces disk I/O bottlenecks; (2) when there are many I/O, it is advantageous to distribute them on several servers. However, each server must have a cache for input data so that it does not compete for the same resource; and (3) GPU-based parallel processing method is most suitable for models such as PRISM with large computation loads.

• Clinical and Experimental Pediatrics
• /
• v.51 no.10
• /
• pp.1102-1111
• /
• 2008

Purpose : Resveratrol, extracted from red wine and grapes, has an anti-cancer effect, an antiinflammatory effect, and an antioxidative effect mainly in heart disease and also has neuroprotective effects in the adult animal model. No studies for neuroprotective effects during the neonatal periods have been reported. Therefore, we studied the neuroprotective effect of resveratrol on hypoxic-ischemic brain damage in neonatal rats via anti-apoptosis. Methods : Embryonic cortical neuronal cell culture of rat brain was performed using pregnant Sprague-Dawley (SD) rats at 18 days of gestation (E18) for the in vitro approach. We injured the cells with hypoxia and administered resveratrol (1, 10, and $30{\mu}g/mL$) to the cells at 30 minutes before hypoxic insults. In addition, unilateral carotid artery ligation with hypoxia was induced in 7-day-old neonatal rats for the in vivo approach. We injected resveratrol (30 mg/kg) intraperitoneally into animal models. Real-time PCR and Western blotting were performed to identify the neuroprotective effects of resveratrol through anti-apoptosis. Results : In the in vitro approach of hypoxia, the expression of Bax, caspase-3, and the ratio of Bax/Bcl-2, indicators of the level of apoptosis, were significantly increased in the hypoxia group compared to the normoxia group. In the case of the resveratrol-treated group, expression was significantly decreased compared to the hypoxia group. And the results in the in vivo approach were the same as in the in vitro approach. Conclusion : The present study demonstrates that resveratrol plays neuroprotective role in hypoxic-ischemic brain damage during neonatal periods through the mechanism of anti-apoptosis.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.167-181
• /
• 2018

Over the past decade, deep learning has been in spotlight among various machine learning algorithms. In particular, CNN(Convolutional Neural Network), which is known as the effective solution for recognizing and classifying images or voices, has been popularly applied to classification and prediction problems. In this study, we investigate the way to apply CNN in business problem solving. Specifically, this study propose to apply CNN to stock market prediction, one of the most challenging tasks in the machine learning research. As mentioned, CNN has strength in interpreting images. Thus, the model proposed in this study adopts CNN as the binary classifier that predicts stock market direction (upward or downward) by using time series graphs as its inputs. That is, our proposal is to build a machine learning algorithm that mimics an experts called 'technical analysts' who examine the graph of past price movement, and predict future financial price movements. Our proposed model named 'CNN-FG(Convolutional Neural Network using Fluctuation Graph)' consists of five steps. In the first step, it divides the dataset into the intervals of 5 days. And then, it creates time series graphs for the divided dataset in step 2. The size of the image in which the graph is drawn is $40(pixels){\times}40(pixels)$, and the graph of each independent variable was drawn using different colors. In step 3, the model converts the images into the matrices. Each image is converted into the combination of three matrices in order to express the value of the color using R(red), G(green), and B(blue) scale. In the next step, it splits the dataset of the graph images into training and validation datasets. We used 80% of the total dataset as the training dataset, and the remaining 20% as the validation dataset. And then, CNN classifiers are trained using the images of training dataset in the final step. Regarding the parameters of CNN-FG, we adopted two convolution filters ($5{\times}5{\times}6$ and $5{\times}5{\times}9$) in the convolution layer. In the pooling layer, $2{\times}2$ max pooling filter was used. The numbers of the nodes in two hidden layers were set to, respectively, 900 and 32, and the number of the nodes in the output layer was set to 2(one is for the prediction of upward trend, and the other one is for downward trend). Activation functions for the convolution layer and the hidden layer were set to ReLU(Rectified Linear Unit), and one for the output layer set to Softmax function. To validate our model - CNN-FG, we applied it to the prediction of KOSPI200 for 2,026 days in eight years (from 2009 to 2016). To match the proportions of the two groups in the independent variable (i.e. tomorrow's stock market movement), we selected 1,950 samples by applying random sampling. Finally, we built the training dataset using 80% of the total dataset (1,560 samples), and the validation dataset using 20% (390 samples). The dependent variables of the experimental dataset included twelve technical indicators popularly been used in the previous studies. They include Stochastic %K, Stochastic %D, Momentum, ROC(rate of change), LW %R(Larry William's %R), A/D oscillator(accumulation/distribution oscillator), OSCP(price oscillator), CCI(commodity channel index), and so on. To confirm the superiority of CNN-FG, we compared its prediction accuracy with the ones of other classification models. Experimental results showed that CNN-FG outperforms LOGIT(logistic regression), ANN(artificial neural network), and SVM(support vector machine) with the statistical significance. These empirical results imply that converting time series business data into graphs and building CNN-based classification models using these graphs can be effective from the perspective of prediction accuracy. Thus, this paper sheds a light on how to apply deep learning techniques to the domain of business problem solving.

• Journal of Intelligence and Information Systems
• /
• v.18 no.3
• /
• pp.185-202
• /
• 2012

.Since the value of information has been realized in the information society, the usage and collection of information has become important. A facial expression that contains thousands of information as an artistic painting can be described in thousands of words. Followed by the idea, there has recently been a number of attempts to provide customers and companies with an intelligent service, which enables the perception of human emotions through one's facial expressions. For example, MIT Media Lab, the leading organization in this research area, has developed the human emotion prediction model, and has applied their studies to the commercial business. In the academic area, a number of the conventional methods such as Multiple Regression Analysis (MRA) or Artificial Neural Networks (ANN) have been applied to predict human emotion in prior studies. However, MRA is generally criticized because of its low prediction accuracy. This is inevitable since MRA can only explain the linear relationship between the dependent variables and the independent variable. To mitigate the limitations of MRA, some studies like Jung and Kim (2012) have used ANN as the alternative, and they reported that ANN generated more accurate prediction than the statistical methods like MRA. However, it has also been criticized due to over fitting and the difficulty of the network design (e.g. setting the number of the layers and the number of the nodes in the hidden layers). Under this background, we propose a novel model using Support Vector Regression (SVR) in order to increase the prediction accuracy. SVR is an extensive version of Support Vector Machine (SVM) designated to solve the regression problems. The model produced by SVR only depends on a subset of the training data, because the cost function for building the model ignores any training data that is close (within a threshold ${\varepsilon}$) to the model prediction. Using SVR, we tried to build a model that can measure the level of arousal and valence from the facial features. To validate the usefulness of the proposed model, we collected the data of facial reactions when providing appropriate visual stimulating contents, and extracted the features from the data. Next, the steps of the preprocessing were taken to choose statistically significant variables. In total, 297 cases were used for the experiment. As the comparative models, we also applied MRA and ANN to the same data set. For SVR, we adopted '${\varepsilon}$-insensitive loss function', and 'grid search' technique to find the optimal values of the parameters like C, d, ${\sigma}^2$, and ${\varepsilon}$. In the case of ANN, we adopted a standard three-layer backpropagation network, which has a single hidden layer. The learning rate and momentum rate of ANN were set to 10%, and we used sigmoid function as the transfer function of hidden and output nodes. We performed the experiments repeatedly by varying the number of nodes in the hidden layer to n/2, n, 3n/2, and 2n, where n is the number of the input variables. The stopping condition for ANN was set to 50,000 learning events. And, we used MAE (Mean Absolute Error) as the measure for performance comparison. From the experiment, we found that SVR achieved the highest prediction accuracy for the hold-out data set compared to MRA and ANN. Regardless of the target variables (the level of arousal, or the level of positive / negative valence), SVR showed the best performance for the hold-out data set. ANN also outperformed MRA, however, it showed the considerably lower prediction accuracy than SVR for both target variables. The findings of our research are expected to be useful to the researchers or practitioners who are willing to build the models for recognizing human emotions.

• Journal of the Korean Academic Society of Industrial Cluster
• /
• v.1 no.1
• /
• pp.67-86
• /
• 2007

Wonju Medical Equipment Industry, despite of its short history, poor sales and weak manpower and so on, have shown remarkable outcomes in a relatively short period. At the end of 2007, totally 79 enterprises (only 4.6% of whole enterprises in Korea) made 10% of the nationwide production and 15% of the nationwide exports with an annual average growth rate of 66.7%, contributing domestic medical equipment industry tremendously. In addition, many leading medical equipment enterprises in various fields already moved or plan to move to Wonju, accelerating Wonju Medical Equipment Cluster. Wonju Medical Equipment Industry Cluster now enters into the growth stage, getting out of the initial business setup stage. Especially, the nomination of Wonju cluster project from the government accelerates networking (e.g. the development of the universal parts, the establishment of the mutual collaboration model among enterprises, and the mutual marketing), making a rapid growth in Wonju Medical Equipment Industry. Wonju Medical Equipment Industry Cluster revealed positive outcomes despite of the weakness in investment size and infra-structure comparing with the other medical industry cluster in the advanced country, while many domestic enterprises pursued their own growth models and thus failed to promote the international competitive power. Wonju Medical Equipment Industry has been developed rapidly. However, there are many challenging problems to support enterprises: small R&D investment and thus weak technology power, difficulties in recruiting R&D engineers, and poor marketing capabilities, financial infrastructure & policies, and network architecture. In order to develop a world-competitive medical equipment industry cluster at Wonju, the complement of infrastructures, the technology innovation, the mutual marketing, and the network expansion to support enterprises are further required. Wonju' s experiences in developing medical equipment industry so far suggest that our own flexible cluster model considering the industry structure and maturity for different regions should be developed, and specific action plans from the local and central governments based on their systematic strategies for industry development should be implemented in order to build world-competitive industry clusters in Korea.