• Proceedings of the Korean Nuclear Society Conference
• /
• 2004.10a
• /
• pp.363-364
• /
• 2004

• The Korean Journal of Applied Statistics
• /
• v.19 no.3
• /
• pp.587-598
• /
• 2006

The number of variables exceeds the number of samples in microarray data. We propose a nonparametric local linear logistic classification procedure using orthogonal components for classifying high-dimensional microarray data. The proposed method is based on the local likelihood and can be applied to multi-class classification. We applied the local linear logistic classification method using PCA, PLS, and factor analysis components as new features to Leukemia data and colon data, and compare the performance of the proposed method with the conventional statistical classification procedures. The proposed method outperforms the conventional ones for each component, and PLS has shown best performance when it is embedded in the proposed method among the three orthogonal components.

• Journal of the Korean Institute of Gas
• /
• v.23 no.4
• /
• pp.19-27
• /
• 2019

The consumption of liquefied natural gas (LNG) has increased annually due to the strengthening of international environmental regulations. In order to produce stable and efficient LNG, it is essential to divide the global (overall) operating condition and construct a quick and accurate monitoring system for each operation condition. In this study, multi-mode monitoring system is proposed to the LNG plant fractionation process. First, global normal operation data is divided to local (subdivide) normal operation data using global principal component analysis (PCA) and k-means clustering method. And then, the data to be analyzed were matched with the local normal mode. Finally, it is determined the state of process abnormality through the local PCA. The proposed method is applied to 45 fault case and it proved to be more than 5~10% efficient compared to the global PCA and univariate monitoring.

• Proceedings of the IEEK Conference
• /
• 2000.07b
• /
• pp.873-876
• /
• 2000

The information-theoretic approach to face recognition is based on the compact coding where face images are decomposed into a small set of basis images. Most popular method for the compact coding may be the principal component analysis (PCA) which eigenface methods are based on. PCA based methods exploit only second-order statistical structure of the data, so higher- order statistical dependencies among pixels are not considered. Independent component analysis (ICA) is a signal processing technique whose goal is to express a set of random variables as linear combinations of statistically independent component variables. ICA exploits high-order statistical structure of the data that contains important information. In this paper we employ the ICA for the efficient feature extraction from face images and show that ICA outperforms the PCA in the task of face recognition. Experimental results using a simple nearest classifier and multi layer perceptron (MLP) are presented to illustrate the performance of the proposed method.

• 한국지구물리탐사학회:학술대회논문집
• /
• 2003.11a
• /
• pp.383-389
• /
• 2003

Clay minerals show a distinct induced-polarization phenomenon, which is one of the most important factors for predicting groundwater flow and contaminant transport. This paper presents a step-by-step process to estimate Cole-Cole parameters from spectral induced-polarization (IP) data measured on the surface of three-dimensional earth. First, the inversion of low-frequency resistivity survey data is made to identify the dc resistivity ${\rho}_dc$ of a volume having IP effects. The other parameters, chargeability m, time constant $\tau$, and frequency dependence c, are sought for the polarizable volume. Next, using multi-frequency data, c can be obtained as high or low asymptotes of the slope of log phase vs. log frequency. Further, for low m, intrinsic $\tau$ is approximated by apparent one, ${\tau}_a$, which is derived from the relation ${{\omega}{\tau}}_a$=1 at an angular frequency $\omega$, where the imaginary component of spectral IP data has an extreme value. Finally, to obtain intrinsic m a two-step linearized procedure has been derived. For a body of given $\tau$ and c, forward modeling with a progression of m values yields a plot of observed vs. intrinsic imaginary components for a frequency. Since this plot is essentially linear, to extract the intrinsic imaginary component is quite simple with an observed value. Using the plot of intrinsic imaginary component vs. m, intrinsic m is determined. We present a synthetic example to illustrate that the Cole-Cole parameters can be recovered from spectral IP data.

• Smart Structures and Systems
• /
• v.31 no.4
• /
• pp.383-392
• /
• 2023

Post-earthquake building condition assessment is crucial for subsequent rescue and remediation and can be automated by emerging computer vision and deep learning technologies. This study is based on an endeavour for the 2nd International Competition of Structural Health Monitoring (IC-SHM 2021). The task package includes five image segmentation objectives - defects (crack/spall/rebar exposure), structural component, and damage state. The structural component and damage state tasks are identified as the priority that can form actionable decisions. A multi-task Convolutional Neural Network (CNN) is proposed to conduct the two major tasks simultaneously. The rest 3 sub-tasks (spall/crack/rebar exposure) were incorporated as auxiliary tasks. By synchronously learning defect information (spall/crack/rebar exposure), the multi-task CNN model outperforms the counterpart single-task models in recognizing structural components and estimating damage states. Particularly, the pixel-level damage state estimation witnesses a mIoU (mean intersection over union) improvement from 0.5855 to 0.6374. For the defect detection tasks, rebar exposure is omitted due to the extremely biased sample distribution. The segmentations of crack and spall are automated by single-task U-Net but with extra efforts to resample the provided data. The segmentation of small objects (spall and crack) benefits from the resampling method, with a substantial IoU increment of nearly 10%.

• Proceedings of the KSRS Conference
• /
• 2007.10a
• /
• pp.468-470
• /
• 2007

This paper is to present an approach for generating coarse resolution (MODIS data) fraction images of forested region in Korea peninsula using forest type area fraction derived from high resolution data (ASTER data) in regional forest area. A 15-m spatial resolution multi-spectral ASTER image was acquired under clear sky conditions on September 22, 2003 over the forested area near Seoul, Korea and was used to select each end-member that represent a pure reflectance of component of forest such as different forest, bare soil and water. The area fraction of selected each end-member and a 500-m spatial resolution MODIS reflectance product covering study area was applied to a linear mixture inversion model for calculating the fraction image of forest component across the South Korea. We found that the area fraction values of each end-member observed from high resolution image data could be used to separate forest cover in low resolution image data.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2005.06a
• /
• pp.690-693
• /
• 2005

In this paper, probabilistic distribution of fatigue life of chassis component is determined statistically by applying the design of experiments and the Pearson system. To construct $p-\varepsilon-N$ curve, the case that fatigue data are random variables is attempted. Probabilistic density function(p.d.f) for fatigue life is obtained by design of experiment and using this p.d.f fatigue reliability about any aimed fatigue life can be calculated. Lower control arm and rear torsion bar of chassis component are selected as examples for analysis. Component load histories, which are obtained by multi-body dynamic simulation for Belsian load history, are used. Finite element analysis are performed using commercial software MSC Nastran and fatigue analysis are performed using FE Fatigue. When strain-life curve itself is random variable, probability density function of fatigue life has very little difference from log-normal distribution. And the case of fatigue data are random variables, probability density functions are approximated to Beta distribution. Each p.d.f is verified by Monte-Carlo simulation.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.133-137
• /
• 2005

In this paper, we investigate visual and quantitative analysis of different tastes in the liquids using multi-array chemical sensor (MACS) based on the ion-selective electrodes (ISEs), which is so called the electronic tongue (E-Tongue) system. We apply the Fuzzy C-means (FCM) algorithm combined with Principal Component Analysis (PCA), which can be used to reduce multi-dimensional data to two- or three-dimensional data, to classify visually data patterns detected by E-Tongue system. The proposed technique can be determined the cluster centers and membership grade of patterns through the unsupervised way. The membership grade of an unknown pattern, which does not shown previously, can be visually and analytically determined. Throughout the experimental trails, the E-tongue system combined with the proposed algorithms is demonstrated robust performance for visual and quantitative analysis for different tastes in the liquids.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.2 no.3
• /
• pp.71-80
• /
• 1999

The aim of the present study is to define and tentatively to interpret the distribution of polluted water released from Lake Sihwa into the Yellow Sea using Landsat TM. Since the region is an extreme Case 2 water, empirical algorithms for detecting concentration of chlorophyll-a and suspended sediments have limitations. This work focuses on the use of multi-temporal Landsat TM data. We applied PCA to detect evolution of spatial feature of polluted water after release from the lake Sihwa. The PCA results were compared with in situ data, such as chlorophyll-a, suspended sediments, Secchi disk depth(SDD), surface temperature, remote sensing reflectance at six channel of SeaWiFS. Also, the in situ remote sensing reflectance obtained by PRR-600(Profiling Reflectance Radiometer) was compared with PCA results of Landsat TM data sets to find good correlation between first Principal Component and Secchi disk depth($R^2$=0.7631), although other variables did not result in such a good correlation. Therefore, Problems in applying PCA techniques to multi-spectral remotely sensed data were also discussed in this paper.