• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.43 no.2
• /
• pp.33-38
• /
• 2020

Anomaly detection of Machine Learning such as PCA anomaly detection and CNN image classification has been focused on cross-sectional data. In this paper, two approaches has been suggested to apply ML techniques for identifying the failure time of big time series data. PCA anomaly detection to identify time rows as normal or abnormal was suggested by converting subjects identification problem to time domain. CNN image classification was suggested to identify the failure time by re-structuring of time series data, which computed the correlation matrix of one minute data and converted to tiff image format. Also, LASSO, one of feature selection methods, was applied to select the most affecting variables which could identify the failure status. For the empirical study, time series data was collected in seconds from a power generator of 214 components for 25 minutes including 20 minutes before the failure time. The failure time was predicted and detected 9 minutes 17 seconds before the failure time by PCA anomaly detection, but was not detected by the combination of LASSO and PCA because the target variable was binary variable which was assigned on the base of the failure time. CNN image classification with the train data of 10 normal status image and 5 failure status images detected just one minute before.

• Biomolecules & Therapeutics
• /
• v.17 no.2
• /
• pp.133-137
• /
• 2009

The root of Astragalus membranaceus is a traditional folk medicine that has been used for many therapeutic purposes in Asia. It reportedly acts as an immunostimulant, tonic, hepatoprotective, diuretic, antidiabetic, analgesic, expectorant, sedative, and anticancer drug. In this study, metabolomic profiling was applied to the roots of A. membranaceus of different ages using NMR coupled with two multivariate statistical analysis methods: such as principal components analysis (PCA) and canonical discriminant analysis (CDA). This allowed various metabolites to be assigned in NMR spectra, including $\gamma$-aminobutyric acid (GABA), aspartic acid, succinic acid, glutamic acid, glutamine, N-acetyl aspartic acid, acetic acid, arginine, alanine, threonine, lactic acid, and valine. The score plot from PCA and also CDA allowed a clear separation between samples according to age.

• Korean Journal of Food Science and Technology
• /
• v.26 no.6
• /
• pp.750-758
• /
• 1994

Flavour components of seven Korean folk sojues, five Chinese kaoliangchiews and Japanese shochu were determined by GC and GC-MS after solid phase extraction with polydivinyl benzene. Less volatile ethyl succinate and ethyl pelargonate were present in Korean folk sojues while volatile ethyl acetate and ethyl butyrate in Chinese kaoliangchiews. In the case of alcohols, the amount of isopentyl alcohol was relatively higher than that of isobutyl alcohol or n-propyl alcohol in Korean folk sojues. On the contrary, less volatile n-propyl alcohol was present more than isopentyl alcohol in Chinese kaoliangchiews. Multivariate statistical analyses involving principal components analysis (PCA) and discriminant analysis (DA) were applied to the GC data. The results of PCA clearly demonstrate that the first principal scores of Korean folk sojues were similar but the second principal scores were different from each other. Classification of Korean sojues and Chinese kaoliangchiews into two groups could be conducted by DA. These results suggested that the common charateristics and identities as a distilled liquors was found in Korean folk sojues.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.38 no.10
• /
• pp.1139-1145
• /
• 2014

Creation of a human skeleton model and characterization of the variation in the bone shape are fundamentally important in many applications of biomechanics. In this paper, we present a parametric shape modeling method for femurs that is based on extracting the main parameter of variations of the femur shape from a 3D model database by using statistical shape analysis. For this shape analysis, principal component analysis (PCA) is used. Application of the PCA to 3D data requires bringing all the models in correspondence to each other. For this reason, anatomical landmarks are used for guiding the deformation of the template model to fit the 3D model data. After subsequent application of PCA to a set of femur models, we calculate the correlation between the dominant components of shape variability for a target population and the anatomical parameters of the femur shape. Finally, we provide tools for visualizing and creating the femur shape using the main parameter of femur shape variation.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2015.07a
• /
• pp.205-207
• /
• 2015

본 논문에서는 먼저 주성분 분석 (Principal components analysis, PCA) 을 활용한 Non-local means (NLM) 을 소개하고, 주성분 분석을 하기 위해 필수적인 공분산 행렬 계산을 효율적으로 하는 방법을 제안한다. NLM 에서의 Neighborhood patch 의 크기를 $S{\times}S=S^2$, 이미지 전체의 픽셀 수를 ${\mathcal{Q}}$ 일 때 공분한 행렬을 계산 하기 위해서는 $S^2{\times}{\mathcal{Q}}$ 크기를 가지는 행렬간의 곱 연산이 필요하다. 결론적으로 본 논문에서는 이 행렬의 크기를 줄임으로써 PSNR (Peak signal-to-noise ratio) 의 손실 없이 NLM 의 복잡도를 줄일 수 있음을 보여준다.

• Journal of Environmental Health Sciences
• /
• v.26 no.4
• /
• pp.1-8
• /
• 2000

The objective of this study was to identify the sources and to estimate the source contributions to the atmospheric TSP(total suspended particulate matter) and PM-10(particulate matter with aerodynamic diameters less than 10$\mu\textrm{m}$) concentration in Taegu area. A total of 84 samples was collected during the January to December 1999. TSP and PM-10 were collected on filters by portable air sampler, and heavy metals in TSP and PM-배 were analyzed by ICO(Inductively Coupled Plasma Spectrometery) after preliminary treatment. The results were follow as : First, annual average of TSP and PM-10 concentration was 123 and 69$\mu\textrm{g}$/㎥ respectively. The concentration of TSP and PM-10 were highest in winter season compared to other seasons. Second, the concentration of Al, Fe, Mn were higher in TSP than in PM-10, indicating that these heavy metals are generally associate with natural contributions. Third, metal combinations showed that a high correlation among concentrations of heavy metals were follows: As Al, Fe and Mn in TSP ; Ni, Cr, Cd and Pb in PM-10. Finally, Statistical analysis was performed using Principal Components Analysis(PCA) in order to find possible sources of the pollutants. The factor analysis was permitted to identify four major sources(soil/road dust resuspension, waste incineration, furl combustion, vehicular emission) in each fraction. These source accounted for at least 83, 85% of variance of TSP and PM-10 concentration in Taegu area.

• Plant Resources
• /
• v.2 no.2
• /
• pp.75-80
• /
• 1999

Twenty two varieties of pecan including wild types were classified based on 6 characters measured by principal component analysis score distance. The results are summarized as fellow. Twenty two varieties were classified into 5 groups based in PCA score distance. Five groups were distinctly characterized by many morphological characters. Total variation could be explained by 51%, 95%, 99% with first, third and fifth principal components respectively. Varimax rotation of the factor loading of the first factors indicated that the first component was highly loaded with leaf characters, the second component with fruit characters, but fruit length was negative loaded. The second, the third and the fourths groups of cultivars had very close genetic parentage similarity.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.24 no.8
• /
• pp.55-61
• /
• 2010

In this paper, we proposed selective extraction method of frequency information and PCA-ELM based diagnosis system for three-phase induction motors. As the first step for diagnosis procedure, DFT is performed to transform the acquired current signal into frequency domain. And then, frequency components are selected according to discriminate order calculated by variance As the next step, feature extraction is performed by principal component analysis (PCA). Finally, we used the classifier based on Extreme Learning Machine (ELM) with fast learning procedure. To show the effectiveness, the proposed diagnostic system has been intensively tested with the various data acquired under different electrical and mechanical faults with varying load.

• Journal of Korean Society of Water and Wastewater
• /
• v.26 no.6
• /
• pp.757-766
• /
• 2012

In this study, we have examined concentration distribution and patterns of PCBs in waters, sediments and soils in an agricultural area of South Korea to investigate the relationship between PCBs sources and concentration levels. The concentration of PCBs in water samples were ranged from lower values below detection limit to 8.25 ug/L and the concentration of PCBs in sediment samples were ranged from lower values below detection limit to 76.67 ug/Kg. The concentration of PCBs in soil samples were ranged from lower values below detection limit to 23.51 ug/Kg. These contamination levels were far below the guideline values suggested for environmental quality assessment. The homologue patterns in samples varied from sample to sample, but isomer patterns were very similar with each other. PCB-138 and PCB-153 were predominant congeners in the soil and sediment, which were similar to the results obtained from previous studies. With these results, the assessment of potential sources of PCBs contamination in the sediments of the Nakdong river basin was performed. The principal components were extracted by Principal Component Analysis(PCA). As the result of PCA, it could be expected that PCBs in samples of this study were more affected by PCB products than combustion processes and mostly affected by already-known sources. The PCBs in the soil and sediment samples were related with commercial PCB products.

• Journal of Korea Multimedia Society
• /
• v.20 no.8
• /
• pp.1299-1311
• /
• 2017

Sounds based machine fault diagnosis recovers all the studies that aim to detect automatically faults or damages on machines using the sounds emitted by these machines. Conventional methods that use mathematical models have been found inaccurate because of the complexity of the industry machinery systems and the obvious existence of nonlinear factors such as noises. Therefore, any fault diagnosis issue can be treated as a pattern recognition problem. We present here an automatic fault diagnosis system of hand drills using discrete wavelet transform (DWT) and pattern recognition techniques such as principal component analysis (PCA) and artificial neural networks (ANN). The diagnosis system consists of three steps. Because of the presence of many noisy patterns in our signals, we first conduct a filtering analysis based on DWT. Second, the wavelet coefficients of the filtered signals are extracted as our features for the pattern recognition part. Third, PCA is performed over the wavelet coefficients in order to reduce the dimensionality of the feature vectors. Finally, the very first principal components are used as the inputs of an ANN based classifier to detect the wear on the drills. The results show that the proposed DWT-PCA-ANN method can be used for the sounds based automated diagnosis system.