• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2003년도 추계 학술발표회 논문집
• /
• pp.69-72
• /
• 2003

The paper considers partial least squares (PLS) as a new dimension reduction technique for the feature vector to overcome the small sample size problem in face recognition. Principal component analysis (PCA), a conventional dimension reduction method, selects the components with maximum variability, irrespective of the class information. So PCA does not necessarily extract features that are important for the discrimination of classes. PLS, on the other hand, constructs the components so that the correlation between the class variable and themselves is maximized. Therefore PLS components are more predictive than PCA components in classification. The experimental results on Manchester and ORL databases show that PLS is to be preferred over PCA when classification is the goal and dimension reduction is needed.

• Communications for Statistical Applications and Methods
• /
• 제22권2호
• /
• pp.173-180
• /
• 2015

We propose a selection procedure of principal components in principal component regression. Our method selects principal components using variable selection procedures instead of a small subset of major principal components in principal component regression. Our procedure consists of two steps to improve estimation and prediction. First, we reduce the number of principal components using the conventional principal component regression to yield the set of candidate principal components and then select principal components among the candidate set using sparse regression techniques. The performance of our proposals is demonstrated numerically and compared with the typical dimension reduction approaches (including principal component regression and partial least square regression) using synthetic and real datasets.

• 한국해양공학회지
• /
• 제34권6호
• /
• pp.469-474
• /
• 2020

This study aims to improve the price competitiveness of a steering system with a relatively high cost portion among sail yacht components. Hence, the components of the steering system were analyzed, and steering system modularization was proposed. The fabrication processes before and after the application of modularization was presented. For modularization, primary components such as the pedestals and quadrants of the steering system were developed, and the structural safety of the components was reviewed. It was confirmed that the manufacturing cost of the developed steering system decreased by approximately 33% compared with the existing system. The new steering system presented herein is expected to contribute to the localization of components and price competitiveness of sail yachts.

• ETRI Journal
• /
• 제40권2호
• /
• pp.275-282
• /
• 2018

Flash reduction of graphene oxide is an efficient method for producing high quality reduced graphene oxide under room temperature ambient conditions without the use of hazardous reducing agents (such as hydrazine and hydrogen iodide). The entire process is fast, low-cost, and suitable for large-scale fabrication, which makes it an attractive process for industrial manufacturing. Herein, we present a simple fabrication method for a flexible in-plane graphene micro-supercapacitor using flash light irradiation. All carbon-based, monolithic supercapacitors with in-plane geometry can be fabricated with simple flash irradiation, which occurs in only a few milliseconds. The thinness of the fabricated device makes it highly flexible and thus useful for a variety of applications, including portable and wearable electronics. The rapid flash reduction process creates a porous graphene structure with high surface area and good electrical conductivity, which ultimately results in high specific capacitance ($36.90mF\;cm^{-2}$) and good cyclic stability up to 8,000 cycles.

• 한국전자통신학회논문지
• /
• 제8권9호
• /
• pp.1299-1305
• /
• 2013

이 논문에서 우리는 MFCC 특징벡터의 차원 저감을 통해 음성 인식에서의 계산량을 줄이는 방법을 조사한다. 특징벡터의 특성분해는 벡터의 성분을 분산의 크기에 따라 배치되도록 선형 변환 시켜준다. 첫 번째 성분은 가장 큰 분산을 가져서 패턴 분류에서 가장 중요한 역할을 한다. 따라서, 분산이 작은 성분들을 제외시키는 차원 저감을 통하여, 계산량을 줄이면서 동시에 음성 인식 성능을 저하시키지 않는 방법을 생각할 수 있다. 실험 결과, MFCC 특징벡터의 성분을 절반 정도로 줄여도 음성인식 오류율에 큰 악영향이 없음이 확인되었다.

• 드라이브 ㆍ 컨트롤
• /
• 제9권4호
• /
• pp.26-31
• /
• 2012

Friction reduction between sliding hydraulic machine components is required to improve efficiency and reliability of hydraulic machineries. It is recently reported that surface texturing on sliding bearing surfaces can reduce the friction force highly. In this paper, numerical analysis is carried out to investigate the effect of dimple numbers and inlet boundary pressures on the lubrication characteristics of a parallel sliding bearing using a commercial computational fluid dynamics (CFD) code, FLUENT. The results show that the pressure distribution, load capacity, dimensionless friction force and leakage with dimple number and their locations, and inlet pressures. The overall lubrication characteristics are highly affected by dimple numbers and boundary pressure. The numerical method adopted and results can be used in design of efficient hydraulic machine components.

• 한국공작기계학회:학술대회논문집
• /
• 한국공작기계학회 2003년도 춘계학술대회 논문집
• /
• pp.117-123
• /
• 2003

This paper deals with pressure ripple and noise reduction characteristics for a hydrostatic transmission(HST) consisting of a variable axial piston pump connected in an open loop to a fixed displacement axial piston motor. Pressure ripples in HST is major source of vibration, which can lead to fatigue failure of components and cause noise. In order to reduce the pressure ripple, an annular tube type hydraulic filter proposes to absorb pressure ripples with the high frequencies components to achieve better noise attenuation in HST. The basic principle tube is applied to propagation of pressure wave, reflection, absorption in cross section of discontinuity and resonance in the hydraulic pipeline. It is experimently confirmed that a hydraulic filter is absorbed to be about 30∼40dB of pressure ripple with high frequencies. These results will assist in modeling and design of noise reduction in hydraulic control systems, and here, should provide a means of designing a quieter HST.

• 한국환경과학회지
• /
• 제24권8호
• /
• pp.1065-1073
• /
• 2015

Urbanization can be significantly affected the hydrologic cycle by increasing flood discharge and heat flux. In order to mitigate these modifications in urban areas, Low Impact Development (LID) technique has been designed and applied in Korea. In order to estimate runoff reduction rate using SWMM LID model, the characteristics of five LID techniques was firstly analyzed for water balance. Vegetated swale and green roof were not reduce flood discharge nor infiltration amount. On the other hand, porous pavement and infiltration trench were captured by infiltration function. The flood reduction rate with LID is substantially affected by their structures and properties, e.g., the percentage of the area installed with LID components and the percentage of the drainage area of the LID components.

• Bulletin of the Korean Chemical Society
• /
• 제16권5호
• /
• pp.392-397
• /
• 1995

o-Phenylenediamine (o-PD) was electropolymerized on glassy carbon electrodes under a potential cycling condition. The resulting polymer films mediated electrons for the reduction of molecular oxygen at pH=1.0. It was found from the RDE, RRDE, and cyclic voltammetry experiments that the modified electrodes reduce oxygen to hydrogen peroxide at about 300 mV lower potential than the bare glassy carbon electrode. The polymer film consisted of more than two components. Among those, only one component was active in oxygen reduction, which was formed mainly in the earlier stage of the electropolymerization. 2,3-Diaminophenazine, a cyclic dimer of o-PD, was also active in the oxygen reduction reaction, from which it was suggested that the active polymeric component has a structural unit similar to the cyclic dimer. Finally, the electropolymerization mechanism for the formation of the active and inactive components has been proposed.

• ETRI Journal
• /
• 제45권1호
• /
• pp.131-137
• /
• 2023

The phase reconstruction process in digital holographic microscopy involves a trade-off between the phase error and the high-spatial-frequency components. In this reconstruction process, if the narrow region of the sideband is windowed in the Fourier domain, the phase error from the DC component will be reduced, but the high-spatial-frequency components will be lost. However, if the wide region is windowed, the 3D profile will include the high-spatial-frequency components, but the phase error will increase. To solve this trade-off, we propose the high-variance pixel averaging method, which uses the variance map of the reconstructed depth profiles of the windowed sidebands of different sizes in the Fourier domain to classify the phase error and the high-spatial-frequency components. Our proposed method calculates the average of the high-variance pixels because they include the noise from the DC component. In addition, for the nonaveraged pixels, the reconstructed phase data created by the spatial frequency components of the widest window are used to include the high-spatialfrequency components. We explain the mathematical algorithm of our proposed method and compare it with conventional methods to verify its advantages.