• Journal of KIISE:Software and Applications
• /
• v.30 no.3_4
• /
• pp.326-338
• /
• 2003

The objective of a neural network design and model selection is to construct an optimal network with a good generalization performance. However, training data include noises, and the number of training data is not sufficient, which results in the difference between the true probability distribution and the empirical one. The difference makes the teaming parameters to over-fit only to training data and to deviate from the true distribution of data, which is called the overfitting phenomenon. The overfilled neural network shows good approximations for the training data, but gives bad predictions to untrained new data. As the complexity of the neural network increases, this overfitting phenomenon also becomes more severe. In this paper, by taking statistical viewpoint, we proposed an integrative process for neural network design and model selection method in order to improve generalization performance. At first, by using the natural gradient learning with adaptive regularization, we try to obtain optimal parameters that are not overfilled to training data with fast convergence. By adopting the natural pruning to the obtained optimal parameters, we generate several candidates of network model with different sizes. Finally, we select an optimal model among candidate models based on the Bayesian Information Criteria. Through the computer simulation on benchmark problems, we confirm the generalization and structure optimization performance of the proposed integrative process of teaming and model selection.

• Progress in Medical Physics
• /
• v.18 no.1
• /
• pp.35-41
• /
• 2007

In this work practical considerations of a pulsed arterial spin labeling MRI are presented to reliable multi-slice perfusion measurements In the human brain. Three parameters were considered in this study. First, In order to improve slice profile and Inversion efficiency of a labeling pulse a high power Inversion pulse of adiabatic hyperbolic secant was designed. A $900^{\circ}$ rotation of the flip angle was provided to make a good slice profile and excellent Inversion efficiency. Second, to minimize contributions of a residual magnetization be4ween Interleaved scans of control and labeling we tested three different conditions which were applied 1) only saturation pulses, 2) only spotter gradients, and 3) combinations of saturation pulses and spotter gradients Applications of bo4h saturation pulses and spoiler gradients minimized the residual magnetization. Finally, to find a minimum gap between a tagged plane and an imaging plane we tested signal changes of the subtracted image between control and labeled Images with varying the gap. The optimum gap was about 20mm. In conclusion, In order to obtain high quality of perfusion Images In human brain It Is Important to use optimum parameters. Before routinely using In clinical studios, we recommend to make optimizations of sequence parameters.

• Journal of Korea Water Resources Association
• /
• v.46 no.1
• /
• pp.73-85
• /
• 2013

The researches on the parameter estimation for storage function method have been conducted for a long time using different methods. However, the determination of the optimal parameters takes a long time and there is a controversy that the proposed optimal parameters do not likely represent the physical characteristics of watershed. In this study, the characteristics of the continuity and storage function equation was analyzed and sensitivities were evaluated. As the result, the only optimal solution is suggested among several local optimums. It is also shown that the lag time is able to be determined using the direct runoff starting time of the watershed. From the sensitivity analysis, it is also proved that the determination of the lag time is very important and the only optimal solution could be found easily after selecting the lag time. Therefore, unlike the traditional optimization method, the proposed method does not take a long time to find the optimal solution which is depending on the characteristics of the rainfall events. The fixed coefficient method which is a method to estimate the optimal parameters of storage function method has been modified using the proposed method. Therefore, the practical efficiency to apply storage function method could be enhanced by applying the proposed method. While the traditional method takes care only the error of the runoff hydrograph, it is very important that the proposed method considers the characteristics of the watershed.

• Korean Journal of Food Science and Technology
• /
• v.50 no.2
• /
• pp.186-190
• /
• 2018

This study was investigated changes of physical properties of soybean sprout by changing single or combinations of parameters in each manufacturing process. In blanching, use of sugar solution and decrease of blanching time reduced hypocotyl diameter and weight loss of soybean sprouts. However, single application among parameters in pre-treatment did not induce significant changes in soybean sprout during cooking of frozen product. The control process was blanching with water, seasoning with salt and sugar, and adding cooked rice with mixing type. Combinations of parameters, which were blanching with sugar solution, seasoning with sugar, and adding cooked rice with topping type, induced significantly higher shear force values ($592{\pm}21g$), larger diameter ($1.58{\pm}0.14mm$), and less weight loss ($13.4{\pm}3.0%$) of soybean sprouts during cooking of frozen product than those from the control process ($498{\pm}24g$, $1.35{\pm}0.13mm$, and $16.0{\pm}1.7%$, respectively) (p<0.05).

• Korean Journal of Food Science and Technology
• /
• v.35 no.6
• /
• pp.1104-1109
• /
• 2003

Response surface methodology (RSM) was used to investigate the effects of the processing parameters on supercritical $CO_2\;(SC-CO_2)$. extraction of total carotenoids and ${\beta}$-cyptoxanthin from Citrus unshiu press cake. The parameters tested were $SC-CO_2$ pressure, dynamic extraction time, and concentration of ethanol added as the modifier to $CO_2$. Experimental data correlated well with the processing parameters (p<0.01), and there was a high statistically significant multiple regression relationship for the extraction of total carotenoids and ${\beta}-cyrptoxanthin$ ($R^2=0.9789$ and 0.9796, respectively). The optimal processing conditions were extraction pressure 33.4 and 37.3 MPa, extraction time 39.6 and 41.0 min, ethanol concentration 18.6 and 17.0% for total carotenoids and ${\beta}-cryptozanthin$, respectively. Maximum extraction yields predicted by RSM were 61.1 and 95.8% ppm, respectively. The extraction yield of total carotenoids increased asymptotically with the increase of the extraction pressure. It increased in proportion to extraction time and concentration of the cosolvent. The extraction yield of ${\beta}-cryptoxanthin$ increased with extraction pressure, extraction time, and concentration of the cosolvent. The extraction time and the concentration of the cosolvent, and the interaction between extraction time and the concentration of the cosolvent significantly affected the extraction yields of carotenoids from C. unshiu press cake.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.9 no.2
• /
• pp.121-131
• /
• 2007

Despite the close linkage in changes between the ecological and hydrological processes in forest ecosystems, an integrative approach has not been incorporated successfully. In this study, based on the vegetation and hydrologic data of the Gwangneung headwater catchment with the Geographic Information System, we attempted such an integrated approach by employing the Regional Hydro-Ecologic Simulation System (RHESSys). To accomplish this, we have (1) constructed the input data for RHESSys, (2) developed an integrated calibration system that enables to consider both ecological and hydrological processes simultaneously, and (3) performed sensitivity analysis to estimate the optimum parameters. Our sensitivity analyses on six soil parameters that affect streamflow patterns and peak flow show that the decay parameter of horizontal saturated hydraulic conductivity $(s_1)$ and porosity decay by depth (PD) had the highest sensitivity. The optimization of these two parameters to estimate the optimum streamflow variation resulted in a prediction accuracy of 0.75 in terms of Nash-Sutcliffe efficiency (NSec). These results provide an important basis for future evaluation and mapping of the watershed-scale soil moisture and evapotranspiration in forest ecosystems of Korea.

• Journal of the Korean Chemical Society
• /
• v.53 no.6
• /
• pp.653-662
• /
• 2009

Developing effective tools for predicting absorption, distribution, metabolism, excretion properties and toxicity (ADME/T) of new chemical entities in the early stage of drug design is one of the most important tasks in drug discovery and development today. As one of these attempts, support vector machines (SVM) has recently been exploited for the prediction of ADME/T related properties. However, two problems in SVM modeling, i.e. feature selection and parameters setting, are still far from solved. The two problems have been shown to be crucial to the efficiency and accuracy of SVM classification. In particular, the feature selection and optimal SVM parameters setting influence each other, which indicates that they should be dealt with simultaneously. In this account, we present an integrated practical solution, in which genetic-based algorithm (GA) is used for feature selection and grid search (GS) method for parameters optimization. hERG ion-channel inhibitor classification models of ADME/T related properties has been built for assessing and testing the proposed GA-GS-SVM. We generated 6 different models that are 3 different single models and 3 different ensemble models using training set - 1891 compounds and validated with external test set - 175 compounds. We compared single model with ensemble model to solve data imbalance problems. It was able to improve accuracy of prediction to use ensemble model.

• Journal of the Microelectronics and Packaging Society
• /
• v.23 no.3
• /
• pp.51-56
• /
• 2016

To fabricate a copper (Cu)-based fine conductive filler having antioxidation property, submicron silver (Ag)-coated Cu particles were fabricated and their antioxidation property was evaluated. After synthesizing the Cu particles of $0.705{\mu}m$ in average diameter by a wet-reduction process, Ag-coated Cu particles were fabricated by successive Ag plating using ethylene grycol solvent. Main process parameters in the Ag plating were the concentration of reductant (ascorbic acid), the injection rate of Ag precursor solution, and the stirring rate in mixed solution. Thus, Ag plating characteristics and the formation of separate fine pure Ag phase were observed with different combinations of process parameters. As a result, formation of the separate pure Ag phase and aggregation between Ag-coated Cu particles could be suppressed by optimization of the process parameters. The Ag-coated Cu particles which were fabricated using optimal conditions showed slight aggregation, but excellent antioxidation property. For example, the particles indicated the weight gain not exceeding 0.1% until $225^{\circ}C$ when they were heated in air at the rate of $10^{\circ}C/min$ and no weight gain until 75 min when they were heated in air at $150^{\circ}C$.

• Journal of Korea Water Resources Association
• /
• v.55 no.7
• /
• pp.545-556
• /
• 2022

It is recommended to use long-term hydrometeorological data for more than the service life of the hydraulic structures and water resource planning. For the purpose of expanding rainfall data, stochastic simulation models, such as Modified Bartlett-Lewis Rectangular Pulse (BLRP) and Neyman-Scott Rectangular Pulse (NSRP) models, have been widely used. The optimal parameters of the model can be estimated by repeatedly comparing the statistical moments defined through a combination of parameters of the probability distribution in the optimization context. However, parameter estimation using relatively small observed rainfall statistics corresponds to an ill-posed problem, leading to an increase in uncertainty in the parameter estimation process. In addition, as shown in previous studies, extreme values are underestimated because objective functions are typically defined by the first and second statistical moments (i.e., mean and variance). In this regard, this study estimated the parameters of the NSRP model using the objective function with the third moment and compared it with the existing approach based on the first and second moments in terms of estimation of extreme rainfall. It was found that the first and second moments did not show a significant difference depending on whether or not the skewness was considered in the objective function. However, the proposed model showed significantly improved performance in terms of estimation of design rainfalls.

• Nuclear Engineering and Technology
• /
• v.54 no.3
• /
• pp.842-848
• /
• 2022

Parametric studies of heat transfer and fluid flow are very important research of interest because the design and operation of fluid flow and heat transfer systems are guided by these parametric studies. The safety of the system operation and system optimization can be determined by decreasing or increasing particular fluid flow and heat transfer parameter while keeping other parameters constant. The parameters that can be varied in order to determine safe and optimized system include system pressure, mass flow rate, heat flux and coolant inlet temperature among other parameters. The fluid flow and heat transfer systems can also be enhanced by the presence of or without the presence of particular effects including gravity effect among others. The advanced Generation IV reactors to be deployed for large electricity production, have proven to be more thermally efficient (approximately 45% thermal efficiency) than the current light water reactors with a thermal efficiency of approximately 33 ℃. SCWR is one of the Generation IV reactors intended for electricity generation. High Performance Light Water Reactor (HPLWR) is a SCWR type which is under consideration in this study. One-eighth of a proposed fuel assembly design for HPLWR consisting of 7 fuel/rod bundles with 9 coolant sub-channels was the geometry considered in this study to examine the effects of system pressure and mass flow rate on wall and fluid temperatures. Gravity effect on wall and fluid temperatures were also examined on this one-eighth fuel assembly geometry. Computational Fluid Dynamics (CFD) code, STAR-CCM+, was used to obtain the results of the numerical simulations. Based on the parametric analysis carried out, sub-channel 4 performed better in terms of heat transfer because temperatures predicted in sub-channel 9 (corner subchannel) were higher than the ones obtained in sub-channel 4 (central sub-channel). The influence of system mass flow rate, pressure and gravity seem similar in both sub-channels 4 and 9 with temperature distributions higher in sub-channel 9 than in sub-channel 4. In most of the cases considered, temperature distributions (for both fluid and wall) obtained at 25 MPa are higher than those obtained at 23 MPa, temperature distributions obtained at 601.2 kg/h are higher than those obtained at 561.2 kg/h, and temperature distributions obtained without gravity effect are higher than those obtained with gravity effect. The results show that effects of system pressure, mass flowrate and gravity on fluid flow and heat transfer are significant and therefore parametric studies need to be performed to determine safe and optimum operating conditions of fluid flow and heat transfer systems.