• Journal of KIISE:Software and Applications
• /
• v.27 no.12
• /
• pp.1202-1210
• /
• 2000

본 논문에서는 효과적인 앙상블 머신의 구축을 위한 새로운 방안을 제시한다. 효과적인 앙상블의 구축을 위해서는 앙상블 멤버들간의 상관관계가 아주 낮아야 하며 또한 각 앙상블 멤버들은 전체 문제를 어느 정도는 정확하게 학습하면서도 서로들간의 불일치 하는 부분이 존재해야 한다는 것이 여러 논문들에 발표되었다. 본 논문에서는 주어진 문제의 다양한 면을 학습한 다수의 앙상블 후보 네트웍을 생성하기 위하여 건설적 학습 알고리즘과 능동 학습 알고리즘을 결합한 형태의 신경망 학습 알고리즘을 이용한다. 이 신경망의 학습은 최소 은닉 노드에서 최대 은닉노드까지 점진적으로 은닉노드를 늘려나감과 동시에 후보 데이타 집합에서 학습에 사용할 훈련 데이타를 점진적으로 선택해 나가면서 이루어진다. 은닉 노드의 증가시점에서 앙상블의 후부 네트웍이 생성된다. 이러한 한 차례의 학습 진행을 한 chain이라 정의한다. 다수의 chain을 통하여 다양한 형태의 네트웍 크기와 다양한 형태의 데이타 분포를 학습한 후보 내트웍들이 생성된다. 이렇게 생성된 후보 네트웍들은 확률적 비례 선택법에 의해 선택된 후 generalized ensemble method (GEM)에 의해 결합되어 최종적인 앙상블 성능을 보여준다. 제안된 알고리즘은 한개의 인공 데이타와 한 개의 실세계 데이타에 적용되었다. 실험을 통하여 제안된 알고리즘에 의해 구성된 앙상블의 최대 일반화 성능은 다른 알고리즘에 의한 그것보다 우수함을 알 수 있다.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1995.10a
• /
• pp.1054-1058
• /
• 1995

There have been many research works for the indirect cutting force measurement in machining process, which deal with the case of one-axis cutting process. In multi-axis cutting process, the main difficulties to estimate the cutting forces occur when the feed direction is reversed. This paper presents the indirect cutting force measurement method in contour NC milling processes by using current signals of servo motors. An artificial neural network (ANN) system are suggested. An artificial neural network(ANN) system is also implemented with a training set of experimental cutting data to measure cutting force indirectly. The input variables of the ANN system are the motor currents and the feedrates of x and y-axis servo motors, and output variable is the cutting force of each axis. A series of experimental works on the circular interpolated contour milling process with the path of a complete circle has been performed. It is concluded that by comparing the ANN system with a dynamometer measuring cutting force directil, the ANN system has a good performance.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.426-428
• /
• 1998

본 논문은 움직임 추정기법 중의 하나인 차영상 분석 기법을 기반으로한 이동 물체 추적 시스템을 제안한다. 실세계와 같은 복잡한 환경에서의 적응성을 높이기 위해 동적인 배경 추출 방법을 제안하고, 이를 바탕으로한 차영상 분석 기법을 이용하여 이동 물체를 탐지한 후 개선된 인공신경망의 경쟁학습 모델인 ART2 학습알고리즘을 이용하여 추적한다. 또한 이동 물체의 평가도 값이 아닌 RGB 컬러정보를 이용한 물체의 특징 벡터를 구한다. 이러한 특징 벡터들은 이동 물체의 모양이나 명암의 변화를 반영한다. 이러한 정보의 변화에 적응성을 갖게 하기위해 개선된 ART2를 사용한다. 그리고 실제 환경에서 보행자를 탐지, 추적하는 실험 결과 Gray 영상보다 정확한 추적이 가능하였다.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.24 no.3
• /
• pp.179-188
• /
• 2012

Performance analyses of vertical breakwaters were conducted for fictitiously designed breakwaters for various water depths to analyze the influence of climate change on the structures. The performance-based design method considering sea level rise and wave height increase due to climate change was used for the performance analysis. One of the problems of the performance-based design method is the large calculation time of wave transformation. To overcome this problem, the SWAN model combined with artificial neural network was used. The significant wave height and principal wave direction at the breakwater site are quickly calculated by using a trained neural network with inputs of deepwater significant wave height and principal wave direction, and tidal level. In general, structural stability becomes low due to climate change impacts, but the trend of stability is different depending on water depth. Outside surf zone, the influence of wave height increase becomes more significant, while that of sea level rise becomes negligible, as water depth increases. Inside surf zone, the influence of both wave height increase and sea level rise diminishes as water depth decreases, but the influence of wave height increase is greater than that of sea level rise. Reinforcement and maintenance policies for vertical breakwaters should be established with consideration of these results.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.33 no.3
• /
• pp.101-109
• /
• 2021

Wave observations using a marine X-band radar are conducted by analyzing the backscattered radar signal from sea surfaces. Wave parameters are extracted using Modulation Transfer Function obtained from 3D wave number and frequency spectra which are calculated by 3D FFT of time series of sea surface images (42 images per minute). The accuracy of estimation of the significant wave height is, therefore, critically dependent on the quality of radar images. Wave observations during Typhoon Maysak and Haishen in the summer of 2020 show large errors in the estimation of the significant wave heights. It is because of the deteriorated radar images due to raindrops falling on the sea surface. This paper presents the algorithm developed to increase the accuracy of wave heights estimation from radar images by adopting convolution neural network(CNN) which automatically classify radar images into rain and non-rain cases. Then, an algorithm for deriving the Hs is proposed by creating different ANN models and selectively applying them according to the rain or non-rain cases. The developed algorithm applied to heavy rain cases during typhoons and showed critically improved results.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.34 no.6
• /
• pp.315-324
• /
• 2022

The experimental data obtained in a wave flume were analyzed using machine learning techniques to establish a model that predicts the input wave height of the wavemaker based on the waves that have experienced wave shoaling and to verify the performance of the established model. For this purpose, artificial neural network (NN), the most representative machine learning technique, and Gaussian process regression (GPR), one of the non-parametric regression analysis methods, were applied respectively. Then, the predictive performance of the two models was compared. The analysis was performed independently for the case of using all the data at once and for the case by classifying the data with a criterion related to the occurrence of wave breaking. When the data were not classified, the error between the input wave height at the wavemaker and the measured value was relatively large for both the NN and GPR models. On the other hand, if the data were divided into non-breaking and breaking conditions, the accuracy of predicting the input wave height was greatly improved. Among the two models, the overall performance of the GPR model was better than that of the NN model.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.2
• /
• pp.541-551
• /
• 1996

This study develops the real time ULTC(Under Load Tap Changer) control strategy with LDC setting values moved dynamically using artificial neural networks. The suggested strategy can improve the ULTC voltage compensation capability by building 2 types of neural networks, ANNs and ANNg. ANNs recognizes the uncompensated MTr sending voltage change caused by the receiving voltage variation. And ANNg dynamically determines the most appropriate ULTC setting valtage chanbe caused by the receiving voltage variation. And ANNg dynamically determines the most appropriate ULTC setting values by recognizing the voltage level obtained from ANNs, and the section load pattern for each time period. In order to evaluate the suggested approach, the ULTC voltage compensation strategy are simulated on a 8 feeder distribution system. Artificial neural networks developed in this study are implemented in FORTRAN language on PC 386.

• KIISE Transactions on Computing Practices
• /
• v.23 no.1
• /
• pp.37-44
• /
• 2017

Games of the AOS genre are classified as an e-sport rather than a recreational computer game. The involved statistical analyses such as game playing patterns and the season's characters gain importance due to the expertise-requiring nature of sports. In this study, the strategic analysis of computer games was conducted by using data mining techniques on League of Legend, a representative AOS game. We designed and tested a winning forecast model using winning percentage prediction techniques such as logistic regression analysis, discriminant analysis, and artificial neural networks. The game data analysis results were represented by a probabilistic graph and used in the visualization tool for game play. Experimental results of the winning forecast model showed a high classification rate of 95% on average with potential for use in establishing various strategies for game play with the visualization tool.

• 제어로봇시스템학회:학술대회논문집
• /
• 1990.10a
• /
• pp.131-134
• /
• 1990

A back-propagation neural network based system for a fault diagnosis of a chemical process is developed. Training data are acquired from FCD(Fault-Consequence Digraph) model. To improve the resolution of a diagnosis, the system is decomposed into 6 subsystems and the training data are composed of 0, 1 and intermediate values. The feasibility of this approach is tested through case studies in a real plant, a naphtha furnace, which has been used to develop a knowledge based expert system, OASYS (Operation Aiding expert SYStem).

• Journal of the Korean Society for Nondestructive Testing
• /
• v.16 no.1
• /
• pp.34-45
• /
• 1996

The signal pattern recognition program which can perform signal acquisition and processing, the extraction and selection of features, the classifier design and the evaluation, is developed and applied to the classification of artificial defects in the weld zone of Austenitic STS304. The neural network classifier is compared with the linear discriminant function classifier and the empirical Bayesian classifier. The signal through a broadband sensor is compared with that through a resonance type sensor. In recognition rate, the neural network classifier is best, and the signal through a broadband sensor is better.