• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권3호
• /
• pp.361-368
• /
• 2004

본 논문에서는 온라인 학습 자료의 비선형 특징(feature) 추출을 위한 새로운 온라인 비선형 주성분분석(OL-NPCA : On-line Nonlinear Principal Component Analysis) 기법을 제안한다. 비선형 특징 추출을 위한 대표적인 방법으로 커널 주성분방법(Kernel PCA)이 사용되고 있는데 기존의 커널 주성분 분석 방법은 다음과 같은 단점이 있다. 첫째 커널 주성분 분석 방법을 N 개의 학습 자료에 적용할 때 N${\times}$N크기의 커널 행렬의 저장 및 고유벡터를 계산하여야 하는데, N의 크기가 큰 경우에는 수행에 문제가 된다. 두 번째 문제는 새로운 학습 자료의 추가에 의한 고유공간을 새로 계산해야 하는 단점이 있다. OL-NPCA는 이러한 문제점들을 점진적인 고유공간 갱신 기법과 특징 사상 함수에 의해 해결하였다. Toy 데이타와 대용량 데이타에 대한 실험을 통해 OL-NPCA는 다음과 같은 장점을 나타낸다. 첫째 메모리 요구량에 있어 기존의 커널 주성분분석 방법에 비해 상당히 효율적이다. 두 번째 수행 성능에 있어 커널 주성분 분석과 유사한 성능을 나타내었다. 또한 제안된 OL-NPCA 방법은 재학습에 의해 쉽게 성능이 항상 되는 장점을 가지고 있다.

• 농업과학연구
• /
• 제47권1호
• /
• pp.67-81
• /
• 2020

The physicochemical parameters of honey are used to determine the botanic origin of honey and to specify the composition criteria for honey in regulations and standards. The parameters of honeydew and blossom honeys from Korean beekeepers were determined to investigate whether they complied with the composition criteria for honey in the food code legislated by Korean authority and to establish the parameters which should be subjected to principal component analysis for improved differentiation of honeys. The fructose and glucose contents of the honeydew honey did not comply with the composition criteria. The ash content of the honey was closely correlated with CIE a^* and CIE L^* The principal component analysis of fructose to glucose ratio, CIE a^*, CIE L^*, ash content, free acidity, and fructose and glucose contents enabled classification of honeydew, chestnut, multifloral, and acacia honeys. Additional advantage of the principal component analysis (PCA) is that the physicochemical parameters, such as fructose to glucose ratio (F/G) and color, can be determined using the analytical instruments for composition criteria and quality control of honey. This study suggested that composition criteria for honeydew honey should be established in the food code in accordance with international standards. The principal component analysis reported in this study resulted in improved classification of the honeys from Korean beekeepers.

• 대한전자공학회논문지SP
• /
• 제42권1호
• /
• pp.9-16
• /
• 2005

본 논문에서는 스테레오 영상에서 좌ㆍ우측 영상을 입력받아 거리 변화에 따른 얼굴인식률을 PCA(Principal Component Analysis) 알고리듬으로 비교한다. 제안된 방법에서는 RGB컬러공간에서 YCbCr컬러공간으로 변환하여 얼굴영역을 검출한다. 또한 스테레오 영상을 이용하여 거리를 취득한 후 추출된 얼굴영상의 확대 및 축소하여 보다 강건한 얼굴영역을 추출하고, PCA 알고리듬으로 인식률을 실험하였다. 취득된 얼굴영상의 평균적인 인식결과로 98.61%(30cm), 98.91%(50cm), 99.05%(100cm), 99.90%(120cm), 97.31%(150cm), 96.71%(200cm)의 인식률을 얻을 수 있었다. 따라서 실험을 통하여 제안된 방법은 거리에 따라 확대 및 축소를 적용하면 높은 인식률을 얻을 수 있음을 보였다.

• 정보과학회 논문지
• /
• 제42권7호
• /
• pp.919-928
• /
• 2015

최근 대용량 데이터 대상의 프라이버시 보호 데이터 마이닝(privacy-preserving data mining: PPDM)이 활발히 연구되고 있다. 본 논문에서는 민감한 데이터 집합의 상관관계를 파악하는데 널리 사용되는 주성분 분석 기반의 PPDM을 다룬다. 일반적으로 주성분 분석은 모든 데이터를 한 곳에 모아 처리해야 하므로 민감한 데이터가 서로에게 공개되고, 상당한 계산량을 요구하며, 또한 데이터를 모으는 과정에서 많은 통신 오버헤드가 발생한다. 이러한 문제를 해결하기 위하여 본 논문은 데이터를 한 곳에 모으지 않고도 주성분 분석을 안전하게 계산하는 효율적인 방법을 제안한다. 제안하는 방법은 노드들 간에 한정된 정보만을 공유하면서도 원래의 주성분 분석 결과와 동일한 결과를 얻을 수 있다. 또한 안전한 주성분 분석에 저차원 변환을 적용하여 안전한 유사 문서 검색에 사용한다. 마지막으로 다양한 실험을 통해 제안한 방법이 대용량의 다차원 데이터에서 효율적으로 동작함을 확인한다.

• 한국멀티미디어학회논문지
• /
• 제20권2호
• /
• pp.134-143
• /
• 2017

Machine fault diagnosis recovers all the studies that aim to detect automatically faults or damages on machines. Generally, it is very difficult to diagnose a machine fault by conventional methods based on mathematical models because of the complexity of the real world systems and the obvious existence of nonlinear factors. This study develops an automatic machine fault diagnosis system that uses pattern recognition techniques such as principal component analysis (PCA) and artificial neural networks (ANN). The sounds emitted by the operating machine, a drill in this case, are obtained and analyzed for the different operating conditions. The specific machine conditions considered in this research are the undamaged drill and the defected drill with wear. Principal component analysis is first used to reduce the dimensionality of the original sound data. The first principal components are then used as the inputs of a neural network based classifier to separate normal and defected drill sound data. The results show that the proposed PCA-ANN method can be used for the sounds based automated diagnosis system.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 학술회의 논문집 정보 및 제어부문 B
• /
• pp.645-648
• /
• 2003

Induction motors are a critical component of industrial processes. Sudden failures of such machines can cause the heavy economical losses and the deterioration of system reliability. Based on the reliability and cost competitiveness of driving system (motors), the faults detection and the diagnosis of system are considered very important factors. In order to perform the faults detection and diagnosis of motors, the vibration monitoring method and motor current signature analysis (MCSA) method are emphasized. In this paper, MCSA method is used for induction motor fault diagnosis. This method analyses the motor's supply current. since this diagnoses faults of the motor. The diagnostic algorithm is based on the principal component analysis(PCA), and the diagnosis system is programmed by using LabVIEW and MATLAB.

• 산업경영시스템학회지
• /
• 제31권3호
• /
• pp.125-130
• /
• 2008

이 논문은 자동차 차체 조립과정에서, 품질관리의 일환으로써, 비접촉 자동측정시스템을 이용하여 검사해야 하는 수많은 비독립적인 검사점을 다변량분산분석과 주성분분석을 이용하여 효율적으로 검사점을 감소시키는 방법을 설명하고 있다. 이 연구의 목적은 다변량분산분석, 주성분 분석의 개념과 이러한 기법들을 산업체 제조분야에서 응용하는 방법을 설명하여 독자의 사례 응용 이해를 돕는데 있으며, 또한 특히 주성분분석을 이용하여 수 많은 비독립적인 검사점을 어떻게 유효하게 줄여나가는지를 보여주고자 한다. 독자의 이해를 돕기 위하여 위와 같은 절차를 순서대로 설명하였으며, 실제 자동차 조립공장에서 발생하는 사례를 수치 예를 들어 설명하였다.

• 한국전산구조공학회:학술대회논문집
• /
• 한국전산구조공학회 2007년도 정기 학술대회 논문집
• /
• pp.27-32
• /
• 2007

Determination of crack depth in filed using the self-calibrating surface wave transmission measurement and the cutting frequency in the transmission function (TRF) is very difficult due to variations of the measurement conditions. In this study, it is proposed to use the measured full TRF as a feature for crack depth assessment. A principal component analysis (PCA) is employed to generate a basis of the measured TRFs for various crack cases. The measured TRFs are represented by their projections onto the most significant principal components. Then artificial neural networks (NNs) using the PCA-compressed TRFs is applied to assess the crack in concrete. Experimental study is carried out for five different crack cases to investigate the effectiveness of the proposed method. Results reveal that the proposed method can be effectively used for the crack depth assessment of concrete structures.

• 대한음성학회지:말소리
• /
• 제54호
• /
• pp.73-83
• /
• 2005

The goal of our research is to build a text-independent speaker identification system that can be used in any condition without any additional adaptation process. The performance of speaker recognition systems can be severely degraded in some unknown mismatched microphone and noise conditions. In this paper, we show that PCA(principal component analysis) can improve the performance in the situation. We also propose an augmented PCA process, which augments class discriminative information to the original feature vectors before PCA transformation and selects the best direction for each pair of highly confusable speakers. The proposed method reduced the relative recognition error by 21%.

• Plant Resources
• /
• 제5권3호
• /
• pp.214-223
• /
• 2002

The objective of this study was to understand the conservation of gene resources and provide information for mass selection' of winter bud characters among the selected populations of Kalopanax septemlobus Koidz using analysis of variance(ANOVA) tests. The obtained results are shown below; 1. Ten populations of K. septemlobus were selected for the study of the variation of winter bud characters in Korea. The results of the analysis of variance(ANOVA) tests shows that there were statistically significant differences in all of the winter bud characters among those populations. 2. Correlation analysis shows that width between Height and DBH(Diameter at breast height) characters have negative relationship with all of the characters, as ABL(Apical branch length), ABW(Apical branch width), AWBL(Apical branch winter bud length), AWBW(Apical branch winter bud width), ABT(Apical branch No. of thorns), ABLB(Apical branch No. of lateral bud) and LBL(Lateral branch length), LBW(Lateral branch width), LBT(Lateral branch No. of thorns), LBLB(Lateral branch No. of lateral bud). 3. The result of principal component analysis(PCA) for winter buds showed that the first principal components(PC' s) to the fourth principal component explains about 78% of the total variation. The first principal component(PC) was correlated with AWBW, LWBW, and LBL and the ratio of ABL/ABW and LBL/LBW out of 16 winter bud characters. The second principal component correlated with ABL, ABW, ABLB, LWBL(Lateral branch winter bud length), and LBW and the ratio of AWBL/AWBW. The third principal component correlated with ABL, ABW, LWBL, LBL, and the ratio of LBL/LBW. The fourth principal component correlated with LBL and the ratio of LWBL/LWBW(Lateral branch winter bud width), LBL/LBW. Therefore, these characters were important to analysis of the variation for winter bud characters among selected populations of K. septemlobus in Korea. 4. Cluster analysis using the average linkage method based on 10 selected populations for the 16 winter bud characters of K. septemlobus in Korea showed a clustering into two groups by level of distance 1.1(Fig. 3). As can be seen in Fig. 3, Group I consisted of three areas(Mt. Sori, Mt. Balwang and Mt. Worak) and Group Ⅱ contisted of seven areas(Suwon, Mt. Chuwang, Mt. Kyeryong, Mt. Kaji, Mt. Jiri, Muan, and Mt. Halla). The result of cluster analysis for winter bud characters corresponded well with principal component analysis, as is shown in Fig. 2.