• Wind and Structures
• /
• v.13 no.1
• /
• pp.57-82
• /
• 2010

The 486m-long roof of Shenzhen Citizens Centre is one of the world's longest spatial lattice roof structures. A comprehensive numerical study of wind effects on the long-span structure is presented in this paper. The discretizing and synthesizing of random flow generation technique (DSRFG) recently proposed by two of the authors (Huang and Li 2008) was adopted to produce a spatially correlated turbulent inflow field for the simulation study. The distributions and characteristics of wind loads on the roof were numerically evaluated by Computational Fluid Dynamics (CFD) methods, in which Large Eddy Simulation (LES) and Reynolds Averaged Navier-Stokes Equations (RANS) Model were employed. The main objective of this study is to explore a useful approach for estimations of wind effects on complex curved roof by CFD techniques. In parallel with the numerical investigation, simultaneous pressure measurements on the entire roof were made in a boundary layer wind tunnel to determine mean, fluctuating and peak pressure coefficient distributions, and spectra, spatial correlation coefficients and probability characteristics of pressure fluctuations. Numerical results were then compared with these experimentally determined data for validating the numerical methods. The comparative study demonstrated that the LES integrated with the DSRFG technique could provide satisfactory prediction of wind effects on the long-span roof with complex shape, especially on separation zones along leading eaves where the worst negative wind-induced pressures commonly occur. The recommended LES and inflow turbulence generation technique as well as associated numerical treatments are useful for structural engineers to assess wind effects on a long-span roof at its design stage.

• Korean Journal of Acupuncture
• /
• v.29 no.3
• /
• pp.431-440
• /
• 2012

Objectives : Recently, network science is very popular topic in various scientific fields and many studies have reported that it gives meaningful results on studying characteristics of a complex system. In this study, based on network theory, we made acupoints network using data of combined acupoints which appeared at "Beijiqianjinyaofang". We focused to find out the distinctive roles of remote and local combinations on the network. Furthermore, we aimed to identify the possibility of numerical and quantitative application to acupuncture researches. Methods : Based on examples of combined acupoints in "Beijiqianjinyaofang", the network consisted of 291 nodes and 2,431 links. The spatial distances between combined acupoints were calculated by the human dummy model. We removed the links step by step for the three cases - remote, local, and random cases, and observed the characteristic changes by calculating path lengths, similarity indices, and clustering coefficients. Also cluster analysis was carried out. Results : The network had a small number of remote links, and a large number of local links. These two links had the distinct characteristics. Whereas the local links formed a cluster of nearby nodes, remote links played a role to increase the correlation between the clusters. Conclusions : These results suggest that acupoints network increases the connectivity between the distal part and the trunk of human body, and enables various combinations of the acupoints. This finding conclusively showed that mechanism of combined acupoints could be interpreted meaningfully by applying network theory in acupuncture researches.

• The Korean Journal of Applied Statistics
• /
• v.28 no.6
• /
• pp.1217-1225
• /
• 2015

Survey research, data is commonly collected through a sample design with complex design features that allow the relative efficiency on the precision of an estimator to be measured using the concept of the design effect compared to simple random sampling as a reference design. This concept is most useful when the design effect can be expressed as a function of various design features. We propose a design effect formula suitable under a stratified multistage sampling by generalizing Gabler et al. (1999, 2006)'s approaches for multistage sampling. Its use can either guide improvement in the design efficiency when in design stage or enable the evaluation of the adopted design features afterwards.

• The KIPS Transactions:PartD
• /
• v.9D no.2
• /
• pp.259-266
• /
• 2002

The main issue in software development is the ability of software project effort and cost estimation in the early phase of software life cycle. The regression models for project effort and cost estimation are presented by function point that is a software sire. The data sets used to conduct previous studies are of ten small and not too recent. Applying these models to 789 project data developed from 1990 ; the models only explain fewer than 0.53 $R^2$(Coefficient of determination) of the data variation. Homogeneous group in accordance with project delivery rate (PDR) divides the data sets. Then this paper presents general effort estimation models using project delivery rate. The presented model has a random distribution of residuals and explains more than 0.93 $R^2$ of data variation in most of PDR ranges.

• Asian-Australasian Journal of Animal Sciences
• /
• v.24 no.8
• /
• pp.1031-1040
• /
• 2011

The objective of this study was to characterize factors influencing genetic improvement of dairy cattle for milk production at farm level. Data were accumulated from 305-day milk yields and pedigree information from 1,921 first-lactation dairy cows that calved from 1990 to 2007 on 161 farms in Central Thailand. Variance components were estimated using average information restricted maximum likelihood procedures. Animal breeding values were predicted by an animal model that contained herd-year-season, calving age, and regression additive genetic group as fixed effects, and cow and residual as random effects. Estimated breeding values from cows that calved in a particular month were used to estimate genetic trends for each individual farm. Within-farm genetic trends (b, regression coefficient of farm milk production per month) were used to classify farms into 3 groups: i) farms with negative genetic trend (b<-0.5 kg/mo), ii) farms with no genetic trend (-0.5 kg/$mo{\leq}b{\leq}0.5$ kg/mo), and iii) farms with positive genetic trend (b>0.5 kg/mo). Questionnaires were used to gather information from individual farmers on educational background, herd characteristics, farm management, decision making practices, and opinion on dairy farming. Farmer's responses to the questionnaire were used to test the association between these factors and farm groups using Fisher's exact test. Estimated genetic trend for the complete population was $0.29{\pm}1.02$ kg/year for cows. At farm level, most farms (40%) had positive genetic trend ($0.63{\pm}4.67$ to $230.79{\pm}166.63$ kg/mo) followed by farms with negative genetic trend (35%; $-173.68{\pm}39.63$ to $-0.62{\pm}2.57$ kg/mo) and those with no genetic trend (25%; $-0.52{\pm}3.52$ to $0.55{\pm}2.68$ kg/mo). Except for educational background (p<0.05), all other factors were not significantly associated with farm group.

• Korean Journal of Agricultural Science
• /
• v.47 no.3
• /
• pp.633-644
• /
• 2020

Rice paper is an element of Vietnamese cuisine that can be used to wrap vegetables and meat. Rice and starch are the main ingredients of rice paper and their mixing ratio is important for quality control. In a commercial factory, assessment of food safety and quantitative supply is a challenging issue. A rapid and non-destructive monitoring system is therefore necessary in commercial production systems to ensure the food safety of rice and starch flour for the rice paper wrap. In this study, fluorescence hyperspectral imaging technology was applied to classify grain flours. Using the 3D hyper cube of fluorescence hyperspectral imaging (fHSI, 420 - 730 nm), spectral and spatial data and chemometric methods were applied to detect and classify flours. Eight flours (rice: 4, starch: 4) were prepared and hyperspectral images were acquired in a 5 (L) × 5 (W) × 1.5 (H) cm container. Linear discriminant analysis (LDA), partial least square discriminant analysis (PLSDA), support vector machine (SVM), classification and regression tree (CART), and random forest (RF) with a few preprocessing methods (multivariate scatter correction [MSC], 1^st and 2^nd derivative and moving average) were applied to classify grain flours and the accuracy was compared using a confusion matrix (accuracy and kappa coefficient). LDA with moving average showed the highest accuracy at A = 0.9362 (K = 0.9270). 1D convolutional neural network (CNN) demonstrated a classification result of A = 0.94 and showed improved classification results between mimyeon flour (MF)1 and MF2 of 0.72 and 0.87, respectively. In this study, the potential of non-destructive detection and classification of grain flours using fHSI technology and machine learning methods was demonstrated.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.40 no.6
• /
• pp.1150-1163
• /
• 2016

This study makes use of meta-analysis to statistically integrate the quantitative results of individual research on the influence of fashion product attributes on purchase; in addition, this study utilizes the effect sizes of correlation coefficients. This study was based on 24 individual studies from January 2000 to March 2015. The meta-analysis analyzed 91 effect sizes and 24 studies and calculated the correlation coefficient effect sizes. A random effect model was employed for meta-analysis because the results for the homogeneity test indicated that the effect sizes of each research were heterogeneous. The analysis results are as follows. First, when the total effect sizes of the fashion product attributes that influence clothing purchase decisions were calculated as .256; this indicated that the effect size is slightly below the mid-sized level. Second, upon examining the effect sizes of the categorized fashion product attributes, the intrinsic attributes yielded .222, extrinsic factors yielded .235, and the compiled attributes yielded .420; this demonstrated that the compiled attributes have a larger effect size than individual attributes. Third, when they were measured by the characteristics of study targets, they were larger for a mixed-gender group than women as well as for ordinary citizens than university students. When the effect sizes were measured based on the characteristics of fashion products as suggested in the study, there were significant differences with respect to sportswear, followed by SPA brand clothing, general clothing, fashion accessories, and designer clothing. The cases where the clothing were purchased in non-retail stores were found to have a slightly larger effect size than those of offline retail stores when the effect sizes were measured based on purchase routes; however, the difference was not statistically significant. Next, an investigation of the trend of effect sizes based on published year via the meta regression analysis indicated slightly larger effect sizes as shown in more recent publications, but this was not statistically significant.

• The korean journal of orthodontics
• /
• v.53 no.5
• /
• pp.336-342
• /
• 2023

Objective: To determine whether the gonial angle on digital panoramic radiographs is associated with vitamin D receptor (VDR) Taql polymorphism. Methods: Genomic DNA samples were collected from the buccal mucosa of patients aged 26-43 years. TaqMan assay for single nucleotide polymorphism genotyping was used to detect the genotype of Taql polymorphism. The gonial angle was measured bilaterally on panoramic radiography. The normal gonial angle was fixed as 121.8°, and it represented the cutoff value for the high gonial angle (HGA) and low gonial angle (LGA) groups. Various genetic models were analyzed, namely dominant (homozygous [AA] vs. heterozygous [AG] + polymorphic [GG]), recessive (AA + AG vs. GG), and additive (AA + GG vs. AG), using the chi-squared test. Results: The reliability of the gonial angle measurement was analyzed using a random sample (26%) of the tests, with the intra-examiner correlation showing an intra-class correlation coefficient of 0.99. The frequencies of the AA, AG, and GG genotypes of rs731236 polymorphism were 40.5%, 41.9%, and 17.6% in the HGA group and 21.8%, 51.0%, and 27.2% in the LGA group, respectively (P = 0.042). A statistically significant difference was observed in the allele frequencies between the two groups (P = 0.011). Moreover, a significant correlation was observed in the dominant genetic model. Conclusions: Taql polymorphism in the VDR gene plays a critical role in the vertical growth of the mandible and decreased gonial angle.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_3
• /
• pp.891-907
• /
• 2023

Compact Advanced Satellite 500-4 (CAS500-4) is scheduled to be launched to collect high spatial resolution data focusing on vegetation applications. To achieve this goal, accurate surface reflectance retrieval through atmospheric correction is crucial. Therefore, a machine learning-based atmospheric correction algorithm was developed to simulate atmospheric correction from a radiative transfer model using Sentinel-2 data that have similarspectral characteristics as CAS500-4. The algorithm was then evaluated mainly for forest areas. Utilizing the atmospheric correction parameters extracted from Sentinel-2 and GEOKOMPSAT-2A (GK-2A), the atmospheric correction algorithm was developed based on Random Forest and Light Gradient Boosting Machine (LGBM). Between the two machine learning techniques, LGBM performed better when considering both accuracy and efficiency. Except for one station, the results had a correlation coefficient of more than 0.91 and well-reflected temporal variations of the Normalized Difference Vegetation Index (i.e., vegetation phenology). GK-2A provides Aerosol Optical Depth (AOD) and water vapor, which are essential parameters for atmospheric correction, but additional processing should be required in the future to mitigate the problem caused by their many missing values. This study provided the basis for the atmospheric correction of CAS500-4 by developing a machine learning-based atmospheric correction simulation algorithm.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.5
• /
• pp.999-1008
• /
• 2023

Early prediction of chronic diseases such as diabetes is an important issue, and improving the accuracy of diabetes prediction is especially important. Various machine learning and deep learning-based methodologies are being introduced for diabetes prediction, but these technologies require large amounts of data for better performance than other methodologies, and the learning cost is high due to complex data models. In this study, we aim to verify the claim that DNN using the pima dataset and k-fold cross-validation reduces the efficiency of diabetes diagnosis models. Machine learning classification methods such as decision trees, SVM, random forests, logistic regression, KNN, and various ensemble techniques were used to determine which algorithm produces the best prediction results. After training and testing all classification models, the proposed system provided the best results on XGBoost classifier with ADASYN method, with accuracy of 81%, F1 coefficient of 0.81, and AUC of 0.84. Additionally, a domain adaptation method was implemented to demonstrate the versatility of the proposed system. An explainable AI approach using the LIME and SHAP frameworks was implemented to understand how the model predicts the final outcome.