• The Journal of the Korea Contents Association
• /
• v.9 no.1
• /
• pp.159-166
• /
• 2009

In this paper, we analyse the noise robustness of MLPs(Multilayer perceptrons) through deriving the probability density function(p.d.f.) of output nodes with additive input noises and the misclassification ratio with the integral form of the p.d.f. functions. Also, we propose linear preprocessing methods to improve the noise robustness. As a preprocessing stage of MLPs, we consider ICA(independent component analysis) and PCA(principle component analysis). After analyzing the noise reduction effect using PCA or ICA in the viewpoints of SNR(Singal-to-Noise Ratio), we verify the preprocessing effects through the simulations of handwritten-digit recognition problems.

• Journal of the Korean Solar Energy Society
• /
• v.39 no.3
• /
• pp.67-79
• /
• 2019

The reliable Typical Meteorological Year (TMY) data, sometimes called Test Reference Year (TRY) data, are necessary in the feasibility study of renewable energy installation as well as zero energy building. In Korea, there are available TMY data; TMY from Korea Institute of Energy Research (KIER), TRY from the Korean Solar Energy Society (KSES) and TRY from Passive House Institute Korea (PHIKO). This study aims at examining their characteristics by using Principle Component Analysis (PCA) at six ground observing stations. First step is to investigate the annual averages of meteorological elements from TMY data and their standard deviations. Then, PCA is done to find which principle components are derived from different TMY data. Temperature and solar irradiance are determined as the main principle component of TMY data produced by KIER and KSES at all stations whereas TRY data from PHIKO does not show similar result from those by KIER and KSES.

• Journal of the Korean Statistical Society
• /
• v.24 no.2
• /
• pp.407-417
• /
• 1995

The h-plot is a graphical technique for displaying the structure of one population's variance-covariance matrix. This follows the mathematical algorithem of the principle component biplot based on the singular value decomposition. But it is known that the singular value decomposition is not resistant, i.e., it is very sensitive to small changes in the input data. In this article, since the mathematical algorithm of the h-plot is equivalent to that of principal component biplot of Choi and Huh (1994), we derive the resistant h-plot.

• Journal of the Korea Computer Industry Society
• /
• v.4 no.4
• /
• pp.479-490
• /
• 2003

CBD is a software development approach based on reusable component and supports easy modification and evolution of software. For the success of this approach, a component must be developed with high cohesion and low coupling. In this paper, we propose the two types of clustering analysis technique based on affinity between use-cases and classes and propose component identification method applying to this technique. We also propose component reference model and CBD methodology framework and perform a ease study to demonstrate how the affinity-based clustering technique is used in component identification method. Component identification method contains three tasks such as component extraction, component specification and component architecting. This method uses object-oriented concept for identifying component, which improves traceability from analysis to implementation and can automatically extract component. This method reflects the low coupling-high cohesion principle for good modularization about reusable component.

• Journal of Korean Society on Water Environment
• /
• v.23 no.6
• /
• pp.814-821
• /
• 2007

This study was estimated water quality to raw water quality management of the Maeri intake station in the Nakdong River using Multivariate Analysis. The results of Principle Component Analysis was explained up to 76.9% of total water quality by three principle components. The 1st, 2nd was explained 44.7%, 17.9% and third was explained 14.3%. Also, the three factors was derived from Factor Analysis. The 1st factor was estimated as the matabolism and organic matter pattern related to algal growth. The 2nd factor was judged as the pollution of pattern related to the discharge from stream of the Nakdong River and 3rd factor was viewed as the hydrological variation pattern related to particle matter. The results of Cluster Analysis were classified into three groups.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.185-189
• /
• 2004

For face recognition system, a face detector which can find exact face region from complex image is needed. Many face detection algorithms have been developed under the assumption that background of the source image is quite simple . this means that face region occupy more than a quarter of the area of the source image or the background is one-colored. Color-based face detection is fast but can't be applicable to the images of which the background color is similar to face color. And the algorithm using neural network needs so many non-face data for training and doesn't guarantee general performance. In this paper, A multi-scale, multi-face detection algorithm using PCA is suggested. This algorithm can find most multi-scaled faces contained in static images with small number of training data in reasonable time.

• The Journal of Fisheries Business Administration
• /
• v.14 no.2
• /
• pp.1-14
• /
• 1983

This paper considers a classification of the 88 islands located at Kyong-nam area in Korea, using by examples of 12 components of the islands. By means of principal component analysis 2 principle components were extracted, which explained a total of 73.7% of the variance. Using an eigen variable criterion (λ>1), no further principle components were discussed. Principal component 1 and 2 explained 63.4% and 10.3% of the total variance respectively, The representation of the unrelated factor scores along the first and second principal axes produced a new information with respect to the classification of the islands. Based upon the representation, 88 islands were classified into 6 groups i. e. A, B, C, D, E, and F according to similarity of the components among them in this paper. The "Group F" belongs to a miscellaneous assortment that does not fit into the logical category. category.

• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.5
• /
• pp.481-486
• /
• 2007

Wireless Sensor Network(WSNs) consists of small sensor nodes with sensing, computation, and wireless communication capabilities. The large number of sensor nodes in a WSN means that there will often be some nodes which give erroneous sensor data owing to several reasons such as power shortage and transmission error. Generally, these sensor data are gathered by a sink node to monitor and diagnose the current environment. Therefore, this can make it difficult to get an effective monitoring and diagnosis. In this paper, to overcome the aforementioned problems, intelligent sensor data validation method based on PCA(Principle Component Analysis) is utilized. Furthermore, a practical implementation using embedded system is given to show the feasibility of the proposed scheme.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.10a
• /
• pp.525-528
• /
• 2004

최근 인터넷 및 무선 통신기술의 광범위한 보급으로 인해 현장 계측 데이터 등과 같은 중요 데이터를 인터넷을 통해 실시간으로 수신 가능케 하는 다양한 형태의 웹 기반 원격 모니터링 시스템이 설계되고 있다. 이러한 웹 모니터링 시스템은 기본적으로 짧은 주기마다 측정된 데이터를 원격의 서버로 전송하는 것이 바람직하나 과도한 통신비 문제로 인해 효율적인 시스템 운영이 어렵다는 문제점을 갖는다. 따라서 본 연구에서는 측정데이터의 변화를 효율적으로 검출할 수 있는 PCA(Principle Component Analysis) 기법과 데이터 압축에 탁월한 특성을 갖는 wavelet 기법을 융합한 새로운 형태의 웹 기반 원격모니터링용 데이터 전송기법을 제안하고 실제 데이터에 적용하여 봄으로써 제안된 기법의 유용성을 확인하고자 한다.

• Journal of Korean Society for Quality Management
• /
• v.13 no.1
• /
• pp.26-30
• /
• 1985

In the light of the properties of colloids, in the surface of disperse phase and dispersion, there exist specific characters such as adsorption or electric double layer, which seems to play important roles in determining the physiochemical properties in the dyeing system. Nylon, wool and silk, the typical amphoteric fibers were dyed with Acid dye and various combinations were prepared by combining pH, temperature and dye concentration, in order to generate flowing electric potential which were measured by microviolt meter and specific conductivity meter. The results were transformed to Zeta potential by Helmholtz-Smoluchowski formular and to surface electric charge density by Suzawa formular, surface dye amount, and effective surface area of fibers, and these data were statistically analysed by principle component analysis.