• Journal of Korea Water Resources Association
• /
• v.56 no.2
• /
• pp.75-89
• /
• 2023

As the Mekong River basin is a nationally shared river, it is difficult to collect precipitation data, and the quantitative and qualitative quality of the data sets differs from country to country, which may increase the uncertainty of hydrological analysis results. Recently, with the development of remote sensing technology, it has become easier to obtain grid-based precipitation products(GPPs), and various hydrological analysis studies have been conducted in unmeasured or large watersheds using GPPs. In this study, rainfall-runoff simulation in the Mekong River basin was conducted using the SWAT model, which is a quasi-distribution model with three satellite GPPs (TRMM, GSMaP, PERSIANN-CDR) and two GPPs (APHRODITE, GPCC). Four water level stations, Luang Prabang, Pakse, Stung Treng, and Kratie, which are major outlets of the main Mekong River, were selected, and the parameters of the SWAT model were calibrated using APHRODITE as an observation value for the period from 2001 to 2011 and runoff simulations were verified for the period form 2012 to 2013. In addition, using the ConvAE, a convolutional neural network model, spatio-temporal correction of original satellite precipitation products was performed, and rainfall-runoff performances were compared before and after correction of satellite precipitation products. The original satellite precipitation products and GPCC showed a quantitatively under- or over-estimated or spatially very different pattern compared to APHPRODITE, whereas, in the case of satellite precipitation prodcuts corrected using ConvAE, spatial correlation was dramatically improved. In the case of runoff simulation, the runoff simulation results using the satellite precipitation products corrected by ConvAE for all the outlets have significantly improved accuracy than the runoff results using original satellite precipitation products. Therefore, the bias correction technique using the ConvAE technique presented in this study can be applied in various hydrological analysis for large watersheds where rain guage network is not dense.

• Journal of the Korean Geotechnical Society
• /
• v.23 no.9
• /
• pp.5-16
• /
• 2007

The simple linear elastic-perfectly plastic model with soil parameters $s_u,\;E_u$ and n of undrained condition is usually applied to predict the displacement of a constructed diaphragm wall(DW) on soft soils during excavation. However, the application of this soil model for finite element analysis could not interpret the continued increment of the lateral displacement of the DW for the large and deep excavation area both during the elapsed time without activity of excavation and after finishing excavation. To study the characteristic behaviors of soil behind the DW during the periods without excavation, a series of tests on soft Bangkok clay samples are simulated in the same manner as stress condition of soil elements happening behind diaphragm wall by triaxial tests. Three kinds of triaxial tests are carried out in this research: $K_0$ consolidated undrained compression($CK_0U_C$) and $K_0$ consolidated drained/undrained unloading compression with periodic decrement of horizontal pressure($CK_0DUC$ and $CK_0UUC$). The study shows that the shear strength of series $CK_0DUC$ tests is equal to the residual strength of $CK_0UC$ tests. The Young's modulus determined at each decrement step of the horizontal pressure of soil specimen on $CK_0DUC$ tests decreases with increase in the deviator stress. In addition, the slope of Critical State Line of both $CK_0UC$ and $CK_0DUC$ tests is equal. Moreover, the axial and radial strain rates of each decrement of horizontal pressure step of $CK_0DUC$ tests are established with the function of time, a slope of critical state line and a ratio of deviator and mean effective stress. This study shows that the results of the unloading compression triaxial tests can be used to predict the diaphragm wall deflection during excavation.

• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.1-19
• /
• 2018

Large amount of data is now available for research and business sectors to extract knowledge from it. This data can be in the form of unstructured data such as audio, text, and image data and can be analyzed by deep learning methodology. Deep learning is now widely used for various estimation, classification, and prediction problems. Especially, fashion business adopts deep learning techniques for apparel recognition, apparel search and retrieval engine, and automatic product recommendation. The core model of these applications is the image classification using Convolutional Neural Networks (CNN). CNN is made up of neurons which learn parameters such as weights while inputs come through and reach outputs. CNN has layer structure which is best suited for image classification as it is comprised of convolutional layer for generating feature maps, pooling layer for reducing the dimensionality of feature maps, and fully-connected layer for classifying the extracted features. However, most of the classification models have been trained using online product image, which is taken under controlled situation such as apparel image itself or professional model wearing apparel. This image may not be an effective way to train the classification model considering the situation when one might want to classify street fashion image or walking image, which is taken in uncontrolled situation and involves people's movement and unexpected pose. Therefore, we propose to train the model with runway apparel image dataset which captures mobility. This will allow the classification model to be trained with far more variable data and enhance the adaptation with diverse query image. To achieve both convergence and generalization of the model, we apply Transfer Learning on our training network. As Transfer Learning in CNN is composed of pre-training and fine-tuning stages, we divide the training step into two. First, we pre-train our architecture with large-scale dataset, ImageNet dataset, which consists of 1.2 million images with 1000 categories including animals, plants, activities, materials, instrumentations, scenes, and foods. We use GoogLeNet for our main architecture as it has achieved great accuracy with efficiency in ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Second, we fine-tune the network with our own runway image dataset. For the runway image dataset, we could not find any previously and publicly made dataset, so we collect the dataset from Google Image Search attaining 2426 images of 32 major fashion brands including Anna Molinari, Balenciaga, Balmain, Brioni, Burberry, Celine, Chanel, Chloe, Christian Dior, Cividini, Dolce and Gabbana, Emilio Pucci, Ermenegildo, Fendi, Giuliana Teso, Gucci, Issey Miyake, Kenzo, Leonard, Louis Vuitton, Marc Jacobs, Marni, Max Mara, Missoni, Moschino, Ralph Lauren, Roberto Cavalli, Sonia Rykiel, Stella McCartney, Valentino, Versace, and Yve Saint Laurent. We perform 10-folded experiments to consider the random generation of training data, and our proposed model has achieved accuracy of 67.2% on final test. Our research suggests several advantages over previous related studies as to our best knowledge, there haven't been any previous studies which trained the network for apparel image classification based on runway image dataset. We suggest the idea of training model with image capturing all the possible postures, which is denoted as mobility, by using our own runway apparel image dataset. Moreover, by applying Transfer Learning and using checkpoint and parameters provided by Tensorflow Slim, we could save time spent on training the classification model as taking 6 minutes per experiment to train the classifier. This model can be used in many business applications where the query image can be runway image, product image, or street fashion image. To be specific, runway query image can be used for mobile application service during fashion week to facilitate brand search, street style query image can be classified during fashion editorial task to classify and label the brand or style, and website query image can be processed by e-commerce multi-complex service providing item information or recommending similar item.

• Journal of Intelligence and Information Systems
• /
• v.24 no.2
• /
• pp.111-124
• /
• 2018

This study aims to develop a classification model for predicting the occurrence of hyperlipidemia, one of the chronic diseases. Prior studies applying data mining techniques for predicting disease can be classified into a model design study for predicting cardiovascular disease and a study comparing disease prediction research results. In the case of foreign literatures, studies predicting cardiovascular disease were predominant in predicting disease using data mining techniques. Although domestic studies were not much different from those of foreign countries, studies focusing on hypertension and diabetes were mainly conducted. Since hypertension and diabetes as well as chronic diseases, hyperlipidemia, are also of high importance, this study selected hyperlipidemia as the disease to be analyzed. We also developed a model for predicting hyperlipidemia using SVM and meta learning algorithms, which are already known to have excellent predictive power. In order to achieve the purpose of this study, we used data set from Korea Health Panel 2012. The Korean Health Panel produces basic data on the level of health expenditure, health level and health behavior, and has conducted an annual survey since 2008. In this study, 1,088 patients with hyperlipidemia were randomly selected from the hospitalized, outpatient, emergency, and chronic disease data of the Korean Health Panel in 2012, and 1,088 nonpatients were also randomly extracted. A total of 2,176 people were selected for the study. Three methods were used to select input variables for predicting hyperlipidemia. First, stepwise method was performed using logistic regression. Among the 17 variables, the categorical variables(except for length of smoking) are expressed as dummy variables, which are assumed to be separate variables on the basis of the reference group, and these variables were analyzed. Six variables (age, BMI, education level, marital status, smoking status, gender) excluding income level and smoking period were selected based on significance level 0.1. Second, C4.5 as a decision tree algorithm is used. The significant input variables were age, smoking status, and education level. Finally, C4.5 as a decision tree algorithm is used. In SVM, the input variables selected by genetic algorithms consisted of 6 variables such as age, marital status, education level, economic activity, smoking period, and physical activity status, and the input variables selected by genetic algorithms in artificial neural network consist of 3 variables such as age, marital status, and education level. Based on the selected parameters, we compared SVM, meta learning algorithm and other prediction models for hyperlipidemia patients, and compared the classification performances using TP rate and precision. The main results of the analysis are as follows. First, the accuracy of the SVM was 88.4% and the accuracy of the artificial neural network was 86.7%. Second, the accuracy of classification models using the selected input variables through stepwise method was slightly higher than that of classification models using the whole variables. Third, the precision of artificial neural network was higher than that of SVM when only three variables as input variables were selected by decision trees. As a result of classification models based on the input variables selected through the genetic algorithm, classification accuracy of SVM was 88.5% and that of artificial neural network was 87.9%. Finally, this study indicated that stacking as the meta learning algorithm proposed in this study, has the best performance when it uses the predicted outputs of SVM and MLP as input variables of SVM, which is a meta classifier. The purpose of this study was to predict hyperlipidemia, one of the representative chronic diseases. To do this, we used SVM and meta-learning algorithms, which is known to have high accuracy. As a result, the accuracy of classification of hyperlipidemia in the stacking as a meta learner was higher than other meta-learning algorithms. However, the predictive performance of the meta-learning algorithm proposed in this study is the same as that of SVM with the best performance (88.6%) among the single models. The limitations of this study are as follows. First, various variable selection methods were tried, but most variables used in the study were categorical dummy variables. In the case with a large number of categorical variables, the results may be different if continuous variables are used because the model can be better suited to categorical variables such as decision trees than general models such as neural networks. Despite these limitations, this study has significance in predicting hyperlipidemia with hybrid models such as met learning algorithms which have not been studied previously. It can be said that the result of improving the model accuracy by applying various variable selection techniques is meaningful. In addition, it is expected that our proposed model will be effective for the prevention and management of hyperlipidemia.

• Journal of Chest Surgery
• /
• v.39 no.3 s.260
• /
• pp.184-193
• /
• 2006

Background: Off-pump coronary artery bypass grafting (OPCAB) has been proven to result in less morbidity. The patients who have left ventricular dysfunction may have benefits by avoiding the adverse effects of the cardiopulmonary bypass. The present study compared early and midterm outcomes of off-pump versus on-pump coronary artery bypass grafting (On pump CABG) in patients with severe left ventricular dysfunction. Material and Method: Ninety hundred forth six patients underwent isolated coronary artery bypass grafting by one surgeon between January 2001 and Febrary 2005.. Data were collected in 100 patients who had left ventricular ejection fraction (L VEF) less than $35\%$ (68 OPCAB; 32 On pump CABG). Mean age of patients were 62.9$\pm$9.0 years in OPCAS group and 63.8$\pm$8.0 years in On pump CABG group. We compared the preoperative risk factors and evaluated early and midterm outcomes. Result: In OPCAB and On pump CABG group, mean number of used grafts per patient were 2.75$\pm$0.72, 2.78$\pm$0.55 and mean number of distal anastomoses were 3.00$\pm$0.79, 3.16$\pm$0.72 respectively. There was one perioperative death in OPCAB group ($1.5\%$). The operation time, ventilation time, ICU stay time, CK-MB on the first postoperative day, and occurrence rate of complications were significantly low in OPCAB group. Mean follow-up time was 26.6$\pm$12.8 months (4${\~}$54 months). Mean LVEF of OPCAB and On pump CABG group improved significantly from $27.1\pm4.5\%$ to $40.7\pm13.0\%$ and $26.9\pm5.4\%$ to $33.3\pm13.7\%$. The 4-year actuarial survival rate of OPCAB and On pump CABG group were $92.2\%,\;88.3\%$ and the 4-year freedom rates from cardiac death were $97.7\%,\;96.4\%$ respectively. There were no significant differences between two groups in 4 year freedom rate from cardiac event and angina. Conclusion: OPCAS improves myocardial function and favors early and mid-term outcomes in patients with severe left ventricular dysfunction compared to On pump CABG group. Therefore, OPCAB is a preferable operative strategy even in patients with severe left ventricular dysfunction.

• Journal of Preventive Medicine and Public Health
• /
• v.29 no.4 s.55
• /
• pp.747-764
• /
• 1996

This study intended to obtain an useful information for health management of lead exposed workers and determine biological monitoring interval in early period of exposure by measuring the lead exposure indices and work duration in all male workers (n=433 persons) exposed less than 1 year in 6 storage battery industries and in 49 males who are not exposed to lead as control. The examined variables were blood lead concentration (PBB), Zinc-protoporphyrin concentration (ZPP), Hemoglobin (HB) and personal history; also measured lead concentration in air (PBA) in the workplace. According to the geometric mean of lead concentration in the air, the factories were grouped into three categories: A; When it is below $0.05mg/m^3$, B; When it is between 0.05 and $0.10mg/m^3$, and C; When it is above $0.10mg/m^3$. The results obtained were as follows: 1. The means of blood lead concentration (PBB), ZPP concentration and hemoglobin(HB) in all male workers exposed to lead less than 1 year in storage battery industries were $29.5{\pm}12.4{\mu}g/100ml,\;52.9{\pm}30.0{\mu}g/100ml\;and\;15.2{\pm}1.1\;gm/100ml$. 2. The means of blood lead concentration (PBB), ZPP concentration and hemoglobin(HB) in control group were $5.8{\pm}1.6{\mu}g/100ml,\;30.8{\pm}12.7{\mu}g/100ml\;and\;15.7{\pm}1.6{\mu}g/100ml$, being much lower than that of study group exposed to lead. 3. The means of blood lead concentration and ZPP concentration among group A were $21.9{\pm}7.6{\mu}g/100,\;41.4{\pm}12.6{\mu}g/100ml$ ; those of group B were $29.8{\pm}11.6{\mu}g/100,\;52.6{\pm}27.9{\mu}g/100ml$ ; those of group C were $37.2{\pm}13.5{\mu}g/100,\;66.3{\pm}40.7{\mu}g/100ml$. Significant differences were found among three factory group(P<0.01) that was classified by the geometric mean of lead concentration in the air, group A being the lowest. 4. The mean of blood lead concentration of workers who have different work duration (month) was as follows ; When the work duration was $1\sim2$ month, it was $24.1{\pm}12.4{\mu}g/100ml$, ; When the work duration was $3\sim4$ month, it was $29.2{\pm}13.4{\mu}g/100ml$ ; and it was $28.9\sim34.5{\mu}g/100ml$ for the workers who had longer work duration than other. Significant differences were found among work duration group(P<0.05). 5. The mean of ZPP concentration of workers who have different work duration (month) was as follows ; When the work duration was $1\sim2$ month, it was $40.6{\pm}18.0{\mu}g/100ml$, ; When the work duration was $3\sim4$ month, it was $53.4{\pm}38.4{\mu}g/100ml$ ; and it was $51.5\sim60.4{\mu}g/100ml$ for the workers who had longer work duration than other. Significant differences were found among work duration group(P<0.05). 6. Among total workers(433 person), 18.2% had PBB concentration higher than $40{\mu}g/100ml$ and 7.1% had ZPP concentration higher than $100{\mu}g/100ml$ ; In workers of factory group A, those were 0.9% and 0.0% ; In workers of factory group B, those were 17.1% and 6.9% ; In workers of factory group C, those were 39.4% and 15.4%. 7. The proportions of total workers(433 person) with blood lead concentration lower than $25{\mu}g/100ml$ and ZPP concentration lower than $50{\mu}g/100ml$ were 39.7% and 61.9%, respectively ; In workers of factory group A, those were 65.5% and 82.3% : In workers of factory group B, those were 36.1% and 60.2% ; In workers of factory group C, those were 19.2% and 43.3%. 8. Blood lead concentration (r=0.177, P<0.01), ZPP concentration (r=0.135, P<0.01), log ZPP (r=0.170, P<0.01) and hemoglobin (r=0.096, P<0.05) showed statistically significant correlation with work duration (month). ZPP concentration (r=0.612, P<0.01) and log ZPP (r=0.614, P<0.01) showed statistically significant correlation with blood lead concentration 9. The slopes of simple linear regression between work duration(month, independent variable) and blood lead concentration (dependent variable) in workplace with low air concentration of lead was less steeper than that of poor working condition with high geometric mean air concentration of lead. The study result indicates that new employees should be provided with biological monitoring including blood lead concentration test and education about personal hygiene and work place management within $3\sim4$ month.

• (The) Research of the performance art and culture
• /
• no.34
• /
• pp.383-421
• /
• 2017

Because the Jitsori and the Hotsori of the Beompae(the Korean Budhhist chant) has no meter and no Jangdan(a Rhythmic cycle of the Korean Music), so it is hard to analyze the melody of the Beompae. Also the melody of the Beompae is different from that of the other Korean traditional music, so studying of the Beompae has been out of the limelight of many scholars, studying the Korean music. But the melody of Beompae had been handed down for thousands of years in Korea, it and other Korean trditional music, had exchanged the impacts each other for a longtime. So I thinks that the Korean Beomapae have shared the similarity of the musical features with the other Korean traditional music. Because the Beompae of the Bulmosan Yeongsanjae on the Geongsangnamdo province has also no meters and no Jangdan, it is difficult to understand, too. But because the Onghoge of Bulmosan Yeongsanjae have a well-regulated melodic structure in comparison with the Beompae of the Seoul province, so called Geongjae Beompae, it seem to be easy to analyze its melody. So I will analyze the melody of Bulmosan Yeongsanjae Onghoge. This analyze should be contribute to investigate the rule of the melodic progress method on the convoluted Beompae melody. Onghoge has been sung on the procedure for Siryeon, Samsiniun(Goebuliun), Jojeonjeoman, Sinjungjakbeop. And the monk for the ritual has sung the chant first to purify the ritual place and to protect the soul. They has called the song, Onghoge a Jitsori at the Bulmosan Yeongsanjae preservation society of the Gyeongnam province. Commonly, there were Jitsori and Hotsori in the Beompae melody, and the melody of Jitsori is longer than that of the Hotsori. So, the melody of Onghoge is lengthened. In other word, the melody of the Onghoge show the lengthened and curved melodic feture of the Beompae very well. Hahn Manyeong, who had studied on the Beompae, Budhhist chant, said that the Hotsori has five letters in a phrase, and there were 4 phrases in a song. And he had insisted that the form of the song, Hotsori, is ABAB. I analyze the melody of the Onghoge by the Hahn's method. I will extract the Wonjeom(a primary tone of a skeletal melodic structure) from the melody of Onghoge, and in the progress of the Wonjeom of Onghoge melodies, I will arrange the repeat of the Wonjeom melody. It is a structural melody of Onghoge. The first phrase of Bulmosan Yeongsanjae Onghoge, 'Pal bu geum gang ho do ryang(八部金剛護道場)' have 4 structural melodies, the second phrase 'Gong sin sog bu bo cheon wang(空神速赴報天王)', the third phrase 'Sam gye je cheon ham le jip(三界諸天咸來集)', the firth phrase 'Yeo geum bul chal bo jeong sang(如今佛刹補禎祥)' have 2 structural melodies each. The structural melodies of Onghoge are 10 in total. And the structural melody of the Onghoge is formed the shape of 'Mi - La - do - La - Mi'. All of the Onghoge melodies is repeated 10 times by the melodic shape. The form of the Onghoge is not ABAB by Hahn, but is 10 times repeat of the shape.