• Journal of Korea Multimedia Society
• /
• v.13 no.4
• /
• pp.487-495
• /
• 2010

It has been required to develop an accurate and reliable evaluation method for the performance of biometric systems as their use is getting popular. Among a number of biometric systems, face recognition is one of the most widely used techniques and this leads to develop a stable evaluation method for face recognition systems in order to standardize the performance of face recognition systems. However, it is considered as a difficult task to evaluation such systems due to a large number of factors that affect their performance. Thus, it may be infeasible to take into account all the environmental factors that are related to the performance of face recognition systems and this naturally suggests an evaluation method for the overall performance based on scenarios. In this paper, we have analyzed environmental factors that are related to the performance of general face recognition systems and proposed their evaluation method taking into account those factors. We have proposed an evaluation method based on scenario that considers the combination of individual environment factors instead of evaluating the performance of face recognition systems regarding each factor. Indeed, we have presented examples on the evaluation of face recognition systems based on scenario that takes into account overall environmental factors.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.41 no.1
• /
• pp.9-19
• /
• 2017

In this study, stator shroud injection in a single-stage transonic axial compressor is proposed. A parametric study of the effect of stator shroud injection on aerodynamic performances was conducted using the three-dimensional Reynolds-averaged Navier-Stokes equations. The curvature, length, width, and circumferential angle of the stator shroud injector and the air injection mass flow rate were selected as the test parameters. The results of the parametric study show that the aerodynamic performances of the single-stage transonic axial compressor were improved by stator shroud injection. The aerodynamic performances were the most sensitive to the injection mass flow rate. Further, the total pressure ratio and adiabatic efficiency were the maximum when the ratio of circumferential angle was 10％.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.35 no.12
• /
• pp.1391-1398
• /
• 2011

Design and performance analysis of a steam turbine for variations of degree of reaction were performed by computer simulation. Design parameters such as blade angles, exit areas, and heights of the nozzle and moving blade were represented as functions of the degree of reaction. The main performance factors such as turbine power, diagram efficiency, and axial thrust were also expressed in terms of the degree of reaction. For further information about the design and performance, the blade angles and main performance factors were investigated as functions of the flow coefficient. The turbine power and diagram efficiency reached a maximum value for a given degree of reaction and flow coefficient, and the symmetric shape of the moving blade showed distortion as the degree of reaction was increased.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2011.05a
• /
• pp.226-226
• /
• 2011

최적 저수지운영을 위한 운영률 도출이나 강우-유출 및 수질 모형의 매개변수 추정 문제처럼 비선형적이고 추정해야할 변수의 수가 많은 경우, 수학적으로 모형화하기에 너무 복잡해서 선형계획법, 비선형계획법, 동적계획법 등을 사용하여 최적해를 구할 수 없는 경우도 있다. 이러한 문제에 대해서는 구조적 진화를 통해 최적해를 구하는 방법들이 사용된다. 일반적으로 미지수의 개수가 많아지면 전역최적해를 찾기가 어려워진다. 전역최적해를 찾는 여러 가지 방법들이 수자원 분야에서는 강우-유출모형의 매개변수를 추정하는데 많이 사용되고 있으며, 특히 유전자 알고리즘, SCE-UA 알고리즘 등 전역최적해를 찾는 메타휴리스틱 방법이 많이 사용되고 있다. 전역최적화 방법을 개발하는 연구자들은 최적화방법의 성능을 평가하기 위해 다양한 검사함수(test function)를 만들어 성능을 평가하고 있다. 본 연구에 사용한 검사함수는 Mishra의 연구(2006a, 2006b)에서 사용한 중요하고 복잡한 검사함수이다. 유전자 알고리즘, SCE-UA 알고리즘, DDS 알고리즘을 검사함수 중 전역해를 찾기 어려운 2 차원 함수 2 가지, 다차원 함수 4 가지 함수에 적용하여 각각의 탐색 성능을 평가하였다. 2차원 함수인 Bukin 함수에 대해서는 모든 최적화 방법이 전역최적해를 찾을 수 없었지만, 유전자 알고리즘이 가장 전역최적해에 가까웠고 다음으로 DDS 알고리즘 순서였다. 지역수렴 영역이 많을 것으로 판단되는 10, 30, 50 차원 Michalewicz 함수에 대해서는 DDS 알고리즘으로 구한 최적해가 전역최적해와 매우 근접하였고 다음으로 SCE-UA 알고리즘, 유전자 알고리즘 순이었다. 지역수렴 영역이 상대적으로 다른 함수보다 넓은 10 차원 Schwefel 함수에 대해서는 DDS 알고리즘으로 구한 최적해가 전역최적해와 거의 근접하였고 유전자 알고리즘과 SCE-UA 알고리즘은 매우 큰 편차를 보였다. 40, 80 차원 Schwefel 함수에 대해서는 3 가지 알고리즘 모두 전역최적해와 편차를 보였지만 DDS 알고리즘에 의한 최적해와 다른 두 알고리즘에 의한 최적해는 1 오더(order) 정도의 차이가 났다. 지역수렴 영역이 큰 Michalewicz 함수와 Schwefel 함수에 대한 결과는 매우 흡사한 결과이다. 이상과 같은 결과로, 유전자 알고리즘은 매개변수의 수가 적을 경우 우수한 탐색성능을 가졌으며, SCE-UA 알고리즘은 Griewank, Rastrigin 함수와 같은 형태인 경우 우수한 성능을 보였다. DDS 알고리즘은 전체적으로 우수한 탐색 능력을 가진 것으로 판단된다. 그러므로 수위구간 영역별 저수지운영률 도출을 위한 적절한 최적화방법으로 DDS 알고리즘을 선정하였다.

• Journal of the Korea Society for Simulation
• /
• v.30 no.1
• /
• pp.75-85
• /
• 2021

In this paper, we present a deep neural network-based prediction model that processes and analyzes the corporate credit and personal credit information of individual business owners as a new method to predict the default rate of individual business more accurately. In modeling research in various fields, feature selection techniques have been actively studied as a method for improving performance, especially in predictive models including many features. In this paper, after statistical verification of macroeconomic indicators (macro variables) and credit information (micro variables), which are input variables used in the default rate prediction model, additionally, through the credit information feature selection method, the final feature set that improves prediction performance was identified. The proposed credit information feature selection method as an iterative & hybrid method that combines the filter-based and wrapper-based method builds submodels, constructs subsets by extracting important variables of the maximum performance submodels, and determines the final feature set through prediction performance analysis of the subset and the subset combined set.

• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.153-166
• /
• 2021

A main goal of pharmacogenomics studies is to predict individual's drug responsiveness based on high dimensional genetic variables. Due to a large number of variables, feature selection is required in order to reduce the number of variables. The selected features are used to construct a predictive model using machine learning algorithms. In the present study, we applied several hybrid feature selection methods such as combinations of logistic regression, ReliefF, TurF, random forest, and LASSO to a next generation sequencing data set of 400 epilepsy patients. We then applied the selected features to machine learning methods including random forest, gradient boosting, and support vector machine as well as a stacking ensemble method. Our results showed that the stacking model with a hybrid feature selection of random forest and ReliefF performs better than with other combinations of approaches. Based on a 5-fold cross validation partition, the mean test accuracy value of the best model was 0.727 and the mean test AUC value of the best model was 0.761. It also appeared that the stacking models outperform than single machine learning predictive models when using the same selected features.

• Journal of Energy Engineering
• /
• v.14 no.3 s.43
• /
• pp.194-202
• /
• 2005

The present study was analytically and experimentally carried out to predict the absorption characteristics on combined heat and mass transfer process in a vertical falling film of variable absorbers. Heat and mass transfer enhancements were analytically investigated. Effects of geometric parameters by insert device (spring) and corrugate, flow pattern on absorption performances has been also investigated. Especially, the optimal values of absorber geometry (ID=22.8mm, L=1150m) and kinetic variables (solution flow rate, flow pattern) for maximum absorption performance has been predicted by the numerical analysis. The maximum absorption performance in a numerical analysis and experiment was shown at the wavy-flow by insert device (spring).

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.5
• /
• pp.917-926
• /
• 2010

It is known that bagging and boosting techniques improve the performance in classification problem. A number of researchers have proved the high performance of bagging and boosting through experiments for categorical response but not for continuous response. We study whether bagging and boosting improve data mining methods for continuous responses such as linear regression, decision tree, neural network through bagging and boosting. The analysis of eight real data sets prove the high performance of bagging and boosting empirically.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.392-395
• /
• 2017

본 연구에서는 일 증발접시 증발량 산정을 위한 딥러닝 (deep learning) 모형의 적용성을 평가하였다. 본 연구에서 적용된 딥러닝 모형은 deep belief network (DBN) 기반 deep neural network (DNN) (DBN-DNN) 모형이다. 모형 적용성 평가를 위하여 부산 관측소에서 측정된 기상자료를 활용하였으며, 증발량과의 상관성이 높은 기상변수들 (일사량, 일조시간, 평균지상온도, 최대기온)의 조합을 고려하여 입력변수집합 (Set 1, Set 2, Set 3)별 모형을 구축하였다. DBN-DNN 모형의 성능은 통계학적 모형성능 평가지표 (coefficient of efficiency, CE; coefficient of determination, $r^2$; root mean square error, RMSE; mean absolute error, MAE)를 이용하여 평가되었으며, 기존의 두가지 형태의 ANN (artificial neural network), 즉 모형학습 시 SGD (stochastic gradient descent) 및 GD (gradient descent)를 각각 적용한 ANN-SGD 및 ANN-GD 모형과 비교하였다. 효과적인 모형학습을 위하여 각 모형의 초매개변수들은 GA (genetic algorithm)를 이용하여 최적화하였다. 그 결과, Set 1에 대하여 ANN-GD1 모형, Set 2에 대하여 DBN-DNN2 모형, Set 3에 대하여 DBN-DNN3 모형이 가장 우수한 모형 성능을 나타내는 것으로 분석되었다. 비록 비교 모형들 사이의 모형성능이 큰 차이를 보이지는 않았으나, 모든 입력집합에 대하여 DBN-DNN3, DBN-DNN2, ANN-SGD3 순으로 모형 효율성이 우수한 것으로 나타났다.

• Journal of the Microelectronics and Packaging Society
• /
• v.5 no.2
• /
• pp.27-36
• /
• 1998

통신시스템의 고발열 전자부품 냉각을 위해 3종류의 써모싸이폰을 각각의 용도에 따라 개 발하였으며 그 각각의 설계변수에 대한 냉각특성을 실험적으로 구하였다. TS-I에서는 증발부 내부 에 금속스크린 메쉬형심지를 삽입함으로써 시간에 따른 온도 변화를 작게 하여 냉각성능 안정성을 확보하였고, TS-II에서는 9W/cm2의 높은 냉각성능을 가진 루프형 써모사이폰을 개발하였으며 TS-III에서는 작동유체의 종류, 파이프개수 와이어 삽입여부등 써모사이폰의 주요 설계변수에 따 른 냉각특성을 구하였다.