• Korean Journal of Veterinary Service
• /
• v.28 no.1
• /
• pp.81-89
• /
• 2005

Recently, mass spectrometry coupled with liquid chromatography (LC/MS) has been a preferred technique for determination of organic compounds in complex matrixes. LC/MS provides a high degree sensitivity and specificity of the compounds of interest. The purpose of this study was to confirm analytical method of residual 6 benzimidazoles (thiabendazole, oxfendazole, mebendazole, albendazole, flubendazole and fenbendazole) in meat by LC/MS. Benzimidazoles were analyzed by LC/MS on XTerra $C_{18}$ column with 0.01% trifluoroacetic acid-acetonitrile (TFA) in a gradient mode as mobile phase, and that were identified by electrospray ionization with selected ion recording mode at 150-350 amu mass range. Residual benzimidazoles were extracted from tissue with ethylacetate, and elute benzimidazoles with $50\%$ acetonitrile. In the LC/MS analysis of benzimidazoles, signal to noise ratio was showed relatively high in the positive mode and special ion in the quality analysis was determined via $[M+H]^+$ and Fragment ions. A spectrum of benzimidazoles was showed from all 6 benzimidazoles

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• Progress in Medical Physics
• /
• v.9 no.4
• /
• pp.201-206
• /
• 1998

IMRT optimization method on multiple slice has been developed by using gradient based algorithm. On about 10-30 CT slices including treatment region of a patient, dose optimization has been performed slice by slice to meet the condition that each organ should be exposed below maximum tolerable doses and that the tumor dose within the range of 100$\pm$5 %. Field size was limited to 8$\times$8 cm$^2$ and in this condition, beam divergence was not taken into account to calculate dose distribution. Total dose distribution was calculated by superposing each beamlet whose dose distribution had been precalculated. In order to investigate beam number dependency, dose optimization was performed for one, three, five, seven, and nine coplanar beams and then each optimization index was evaluated. It is found that optimization time was proportional to number of slices to be optimized, and the most efficient plan was obtained from the case of three-to-seven incident beams with respect to calculation time and optimization index. In conclusion, dose optimization of multiple slice was able to be obtained by repeating dose optimization of single slice under condition that the beam size is not too large to ignore beam divergence. And it turns out that result of dose optimization was so sensitive to the position of isocenter that some method to optimize isocenter position is needed to improve it.

• Journal of Korea Water Resources Association
• /
• v.54 no.12
• /
• pp.1255-1263
• /
• 2021

While it could become an alternative water resource, fog could undermine traffic safety and operational performance of infrastructures. To reduce such adverse impacts, it is necessary to have spatially continuous fog risk information. In this work, tree-based machine-learning models were developed in order to quantify fog risks with routine meteorological observations alone. The Extreme Gradient Boosting (XGB), Light Gradient Boosting (LGB), and Random Forests (RF) were chosen for the regional fog models using operational weather and visibility observations within the Jeollabuk-do province. Results showed that RF seemed to show the most robust performance to categorize between fog and non-fog situations during the training and evaluation period of 2017-2019. While the LGB performed better than in predicting fog occurrences than the others, its false alarm ratio was the highest (0.695) among the three models. The predictability of the three models considerably declined when applying them for an independent period of 2020, potentially due to the distinctively enhanced air quality in the year under the global lockdown. Nonetheless, even in 2020, the three models were all able to produce fog risk information consistent with the spatial variation of observed fog occurrences. This work suggests that the tree-based machine learning models could be used as tools to find locations with relatively high fog risks.

• Atmosphere
• /
• v.30 no.4
• /
• pp.377-390
• /
• 2020

Not only emissions, but also atmospheric circulation is a key factor that affects local particulate matters (PM) concentrations in Korea through ventilation effects and transboundary transports. As part of the atmospheric circulation, air stagnation especially adversely affects local air quality due to weak ventilation. This study investigates the large-scale circulation related to air stagnation over Korea during winter and projects the climate change impacts on atmospheric patterns, using observed PM data, reanalysis and regional climate projections from HadGEM3-RA with Modified Korea Particulate matter Index. Results show that the stagnation affects the PM concentration, accompanied by pressure ridge at upper troposphere and weaken zonal pressure gradient at lower troposphere. Downscaling using HadGEM3-RA is found to yield Added-Value in the simulated low tropospheric winds. For projection of future stagnation, SSP5-8.5 and SSP1-2.6 (high and low emission) scenarios are used here. It has been found that the stagnation condition occurs more frequently by 11% under SSP5-8.5 and by 5% under SSP1-2.6 than in present-day climate and is most affected by changes in surface wind speed. The increase in the stagnation conditions is related to anticyclonic circulation anomaly at upper troposphere and weaken meridional pressure gradient at lower troposphere. Considering that the present East Asian winter monsoon is mainly affected by change in zonal pressure gradient, it is worth paying attention to this change in the meridional gradient. Our results suggest that future warming condition increase the frequency of air stagnation over Korea during winter with response of atmospheric circulation and its nonlinearity.

• Journal of Korean Society of Forest Science
• /
• v.100 no.2
• /
• pp.202-211
• /
• 2011

The geometric characteristics of step-pool structures and how they are influenced by channel characteristics were investigated in a steep mountain stream in the Experimental Forests of Kangwon National University in Chuncheon, Gangwon-do. Average values of steps for the study reaches were as follows: step spacing, 4.69 m; step height, 0.47 m; step drop, 0.71 m; step-forming particle sizes, 0.68 m; number, 21steps/ 100 m; the ratio of step spacing to channel width, 0.5; and step steepness, 0.13. Relationships between spacing and height of steps and channel gradient showed a negative- and positive correlation, respectively, whereas all geometric variables of steps manifested poor correlation with channel width. Therefore, step steepness, expressed as the ratio of step height to step spacing, increased as channel gradient increased. The ratio of step steepness to channel gradient representing the criterion of maximum flow resistance was 1.2, indicating the channel bed's stable condition. In particular, the relationship between the ratio of step drop to step height and channel gradient showed a significant negative correlation, suggesting the influence of step-pool geometry in trapping sediment and providing an aquatic habitat. Positive correlations also exist between spacing and drop of steps and step particles. Our findings suggest that the dynamics of step-pool structures may strongly control physical and ecological environments in steep mountain streams, so understanding them is essential for stream management.

• The Korean Journal of Ecology
• /
• v.23 no.6
• /
• pp.439-444
• /
• 2000

Forest vegetation in Mt. Jangan County Park, Jeonlabuk-Do, Korea, was investigated by classification and ordination methods. By the cluster analysis (classification) method, nine groups were recognized as follows : Quercus serrata community, Quercus serrata- Carpinus laxiflora community, Cornus controversa community, Fraxinus mandshurica community, Carpinus laxiflora community, Quereus variabilis community, Quercus mongolica - Sasa borealis community. Quercus mongolica - Symplocos chinensis for. pilosa community and Quercus mongolica - Rhododendron schlippenbachii community. These groups showed differences in species composition and environmental characteristics, but Quercus mongolica - Sasa borealis community, Quercus mongolica - Symplocos chinensis for. pilosa community and Quercus mongolica - Rhododendron schlippenbachii community among them showed very similar floristic composition to each other. The interrelationship between the floristic composition of the vegetation and environmental factors was analysed by principal component analysis (PCA). Quercus mongolica community was distributed at a high altitude (900～1200 m above sea level). Fraxinus mandshurica community and Cornus controversa community were differentiated from the other communities with high contents of soil moisture and pH. On the other hand, Carpinus laxiflora community and Quercus variabilis community were distributed at places with adequate levels of soil moisture, soil organic matter. and at low altitude. In this study, the altitude and soil moisture were the main factors determining the forest vegetation. They were strongly correlated with the dominant compositional gradient at the localities examined.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.22 no.4
• /
• pp.239-249
• /
• 2020

Rice flour varieties have been developed to replace wheat, and consumption of rice flour has been encouraged. damage related to pre-harvest sprouting was occurring due to a weather disaster during the ripening period. Thus, it is necessary to develop pre-harvest sprouting rate prediction system to minimize damage for pre-harvest sprouting. Rice cultivation experiments from 20 17 to 20 19 were conducted with three rice flour varieties at six regions in Gangwon-do, Chungcheongbuk-do, and Gyeongsangbuk-do. Survey components were the heading date and pre-harvest sprouting at the harvest date. The weather data were collected daily mean temperature, relative humidity, and rainfall using Automated Synoptic Observing System (ASOS) with the same region name. Gradient Boosting Machine (GBM) which is a machine learning model, was used to predict the pre-harvest sprouting rate, and the training input variables were mean temperature, relative humidity, and total rainfall. Also, the experiment for the period from days after the heading date (DAH) to the subsequent period (DA2H) was conducted to establish the period related to pre-harvest sprouting. The data were divided into training-set and vali-set for calibration of period related to pre-harvest sprouting, and test-set for validation. The result for training-set and vali-set showed the highest score for a period of 22 DAH and 24 DA2H. The result for test-set tended to overpredict pre-harvest sprouting rate on a section smaller than 3.0 %. However, the result showed a high prediction performance (R2=0.76). Therefore, it is expected that the pre-harvest sprouting rate could be able to easily predict with weather components for a specific period using machine learning.

• Communications for Statistical Applications and Methods
• /
• v.21 no.5
• /
• pp.423-433
• /
• 2014

Test statistics using cumulative sums of residuals have been widely used in various regression models including generalized linear models(GLM). Recently, Pan and Lin (2005) extended this testing procedure to the generalized linear mixed models(GLMM) having random effects, in which we encounter difficulties in computing the marginal likelihood that is expressed as an integral of random effects distribution. The Gaussian quadrature algorithm is commonly used to approximate the marginal likelihood. Many commercial statistical packages provide an option to apply this type of goodness-of-fit test in GLMs but available programs are very rare for GLMMs. We suggest a computational algorithm to implement the testing procedure in GLMMs by a freely accessible R package, and also illustrate through practical examples.

• Journal of Advanced Marine Engineering and Technology
• /
• v.24 no.2
• /
• pp.50-56
• /
• 2000

This paper presents the design of an optimal state feedback controller for container cranes under some design specifications. To do this, the nonlinear equation of a container crane system is linearized and then augmented to eliminate the steady-state error, and some constraints are derived from the design specifications. Designing the controller involves a constrained optimization problem which classical gradient-based methods have difficulties in handling. Therefore, a real-coding genetic algorithm incorporating the penalty strategy is used. The responses of the proposed control system are compared with those of the unconstrained optimal control system to illustrate the efficiency.