• Atmosphere
• /
• v.24 no.4
• /
• pp.565-571
• /
• 2014

A long-range prediction system of tropical cyclone (TC) activity over the western North Pacific (WNP) has been operated in the National Typhoon Center of the Korea Meteorological Administration since 2012. The model forecasts the spatial distribution of TC tracks averaged over the period June~October. In this study, we separately developed TC prediction models for summer (June~August) and autumn (September~November) period based on the current operating system. To perform the three-month WNP TC activity prediction procedure readily, we modified the shell script calling in environmental variables automatically. The user can apply the model by changing these environmental variables of namelist parameter in consideration of their objective. The validations for the two seasons demonstrate the great performance of predictions showing high pattern correlations between hindcast and observed TC activity. In addition, we developed a post-processing script for deducing TC activity in the Korea emergency zone from final forecasting map and its skill is discussed.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.11
• /
• pp.21-30
• /
• 2024

High-dimensional data causes difficulties in machine learning due to high time consumption and large memory requirements. In particular, in a multi-label environment, higher complexity is required as much as the number of labels. This paper proposes a feature selection method to improve classification performance in multi-label settings. The method considers three types of relationships: between features, between features and labels, and between labels themselves. To achieve this, a regression-based objective function is designed. This objective function calculates the linear relationships between features and labels and uses mutual information to compute relationships between features and between labels. By minimizing this objective function, the optimal weights for feature selection are found. To optimize the objective function, a gradient descent method is applied to develop a fast-converging algorithm. The experimental results on six multi-label datasets show that the proposed method outperforms existing multi-label feature selection techniques. The classification performance of the proposed method, averaged over six datasets, showed a Hamming loss of 0.1285, a ranking loss of 0.1811, and a multi-label accuracy of 0.6416. Compared to the AMI(Approximating Mutual Information) algorithm, the performance was better by 0.0148, 0.0435, and 0.0852, respectively.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.12
• /
• pp.603-610
• /
• 2016

Over 12 years have passed since the first commercial operation of a Korean high-speed train. Since then, the transport capacity of the high-speed lines has become almost saturated. Therefore, studies have been carried out to increase the operating speed of the trains in order to increase their transportation capacity. This study was carried out to improve the critical speed of the KTX-Sancheon, Korean high-speed train, in order to increase its operating speed. A dynamic analysis of the KTX-Sancheon train was performed using the contact data obtained from the wheel wear profiles that were measured from a KTX-Sancheon train in commercial operation. The analysis results were verified by comparing them with the measurement acceleration data obtained from KTX-Sancheon. The suspension parameters were optimized to improve the operation speed. The critical speed of KTX-Sancheon was increased by 9.4% after the optimization by the response surface method. The optimized suspension parameters are expected to be used for the new bogie design to increase the operating speed of KTX-Sancheon from 300km/h to 350km/h.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.31 no.10
• /
• pp.51-59
• /
• 2003

A method of autonomous update is presented for onboard orbit propagator. On board propagator is an alternative means that could be used for navigation purpose in case of CPS receiver's failure. Although the ground station is not a able to upload a new propagator, the onboard propagator must be maintained most up-to-date. For this, a filtering technique is proposed wherein GPS data are effectively used to continuously update the on board propagator which was uploaded previously. Even if the ground station has generated the on board propagator based on the wrong information, the onboard propagator with updating scheme can automatically correct the errors in the coefficients of residual reconstruction function. Several scenarios were used to show the validity of the scheme for updating the onboard propagator using KOMPSAT-1 orbit data.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.20 no.6
• /
• pp.29-37
• /
• 2016

This paper proposes a shape optimization of the metal boss for a composite motor case using finite element analysis. For the structural safety and the weight reduction of the composite motor case, under the internal pressure, the fiber stress in the dome area and the tightening bolt stress are constrained and the boss weight is set to objective function, respectively. The response surface models are constructed for the performance characteristics by using response surface method. The significance of the design variables about the performance characteristics is evaluated through the ANOVA(analysis of variance) and the goodness of fit test for the constructed model is performed through the regression analysis. The SQP(sequential quadratic programming) algorithm is used for the optimization and the proposed method is verified by performing structural analysis for the optimum shape.

• Resources Recycling
• /
• v.28 no.2
• /
• pp.23-30
• /
• 2019

An color sorting technique was introduced to recycle End-of-life automobile shredded bumpers. The color sorting is a innovate method of separating the differences in the color of materials which are difficult to separate in gravity and size classification by using a camera and an image process technique. Experiments were planned and optimal conditions were derived by applying BBD (Box-Behnken Design) in the reaction surface method. The effects of color sensitivity, feed rate and sample size were analyzed, and a second-order reaction model was obtained based on the analysis of regression and statistical methods and $R^2$ and p-value were 99.56% and < 0.001. Optimum recovery was 94.1% under the conditions of color sensitivity, feed rate and particle size of 32%, 200 kg/h, and 33 mm respectively. The recovery of actual experiment was 93.8%. The experimental data agreed well with the predicted value and confirmed that the model was appropriate.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.10
• /
• pp.1469-1476
• /
• 2022

Recently, many researches have been conducted to estimate blood pressure using ECG(Electrocardiogram) and PPG(Photoplentysmography) signals. In this paper, we designed and implemented a mobile system to monitor blood pressure in real time by using 1-D convolutional neural networks. The proposed model consists of deep 11 layers which can learn to extract various features of ECG and PPG signals. The simulation results show that the more the number of convolutional kernels the learned neural network has, the more detailed characteristics of ECG and PPG signals resulted in better performance with reduced mean square error compared to linear regression model. With receiving measurement signals from wearable ECG and PPG sensor devices attached to the body, the developed system receives measurement data transmitted through Bluetooth communication from the devices, estimates systolic and diastolic blood pressure values using a learned model and displays its graph in real time.

• Journal of the Society of Cosmetic Scientists of Korea
• /
• v.48 no.3
• /
• pp.255-264
• /
• 2022

The active ingredients of Napttegi Kong(GML, Glycine max landrace), a type of native rare soybeans, were identified, and an ultrasonic extraction method was introduced as an eco-friendly extraction method. Through the component analysis of the Napttegi Kong extract, the epicatechin, which was not found in conventional soybeans, was identified. For effective extraction using ultrasonic, the main extraction conditions were optimized using the response surface analysis method. Through the Box-Behnken design process, 15 experiments were conducted with the extraction temperature, the ratio of extraction solvent/solution, and extraction time as key independent variables. A quadratic regression equation for the two dependent variables, epicatechin content and total isoflavone content, was derived, and the coefficients of determination were found to be high as R² = 0.9939 and R² = 0.9844, respectively, confirming that the correlation showed high significance. The extraction conditions satisfying the maximum expectations of these two dependent variables were predicted. to be 40.4℃ of extraction temperature, 19.3 times of extraction solvent/solution, and 91 sec of extraction time. The expected value and the actual experimental value of the epikatechin content and the total isoflavone content were similar, so it was confirmed that this experimental method is a highly reliable optimization model.

• Applied Chemistry for Engineering
• /
• v.34 no.2
• /
• pp.137-143
• /
• 2023

In this study, natural surfactants were extracted from Medicago sativa L. The O/W emulsification processes with the extracted natural surfactants were optimized using central composite design model-response surface methodology (CCD-RSM) and a 95% confidence interval was used to confirm the reasonableness of the optimization. Herein, independent parameters were the ratio of saponins to total surfactant (P), amount of surfactant (W), and emulsification speed (R), whereas the reaction parameters were the emulsion stability index (ESI), mean droplet size (MDS), and viscosity (V). Using the multiple reaction, the optimal conditions for the ratio of saponins to total surfactant, amount of surfactant, and emulsification speed for O/W emulsification were 49.5%, 9.1 wt%, and 6559.5 rpm, respectively. Under these optimal conditions, the expected values of ESI, MDS, and V as the reaction parameters were 89.9%, 1058.4 nm, and 1522.5 cP, respectively. The values of ESI, MDS, and V from these expected values were 88.7%, 1026.4 nm, and 1486.5 cP, respectively, and the average experimental error for validating the accuracy was about 2.3 (± 0.4)%. Therefore, it was possible to design an optimization process for evaluating the O/W emulsion process with Medicago sativa L. using CCD-RSM.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.127-137
• /
• 2019

Recidivism prediction has been a subject of constant research by experts since the early 1970s. But it has become more important as committed crimes by recidivist steadily increase. Especially, in the 1990s, after the US and Canada adopted the 'Recidivism Risk Assessment Report' as a decisive criterion during trial and parole screening, research on recidivism prediction became more active. And in the same period, empirical studies on 'Recidivism Factors' were started even at Korea. Even though most recidivism prediction studies have so far focused on factors of recidivism or the accuracy of recidivism prediction, it is important to minimize the prediction misclassification cost, because recidivism prediction has an asymmetric error cost structure. In general, the cost of misrecognizing people who do not cause recidivism to cause recidivism is lower than the cost of incorrectly classifying people who would cause recidivism. Because the former increases only the additional monitoring costs, while the latter increases the amount of social, and economic costs. Therefore, in this paper, we propose an XGBoost(eXtream Gradient Boosting; XGB) based recidivism prediction model considering asymmetric error cost. In the first step of the model, XGB, being recognized as high performance ensemble method in the field of data mining, was applied. And the results of XGB were compared with various prediction models such as LOGIT(logistic regression analysis), DT(decision trees), ANN(artificial neural networks), and SVM(support vector machines). In the next step, the threshold is optimized to minimize the total misclassification cost, which is the weighted average of FNE(False Negative Error) and FPE(False Positive Error). To verify the usefulness of the model, the model was applied to a real recidivism prediction dataset. As a result, it was confirmed that the XGB model not only showed better prediction accuracy than other prediction models but also reduced the cost of misclassification most effectively.