• 한국해양공학회:학술대회논문집
• /
• 한국해양공학회 2001년도 추계학술대회 논문집
• /
• pp.232-237
• /
• 2001

The evaluation of specimen thickness effect of fatigue crack growth life by the simulation of probabilistic fatigue crack growth is presented. In this paper, the material resistance to fatigue crack growth is treated as a spatial stochastic process, which varies randomly on the crack surface. Using the previous experimental data, the non-Gaussian(eventually Weibull, in this report) random fields simulation method is applied. This method is useful to estimate the probability distribution of fatigue crack growth life and the variability due to specimen thickness by simulating material resistance to fatigue crack growth along a crack path.

• 한국전자통신학회논문지
• /
• 제7권4호
• /
• pp.827-832
• /
• 2012

본 논문에서는 주기 $2^n-1$인 m-수열에 새로운 데시메이션을 적용하여 얻은 Gold 계열의 이진수열을 제안하고, 제안된 이진수열의 상호상관관계 함숫값이 4개임을 보인다.

• Communications for Statistical Applications and Methods
• /
• 제19권1호
• /
• pp.193-211
• /
• 2012

This paper deals with robust nonparametric regression analysis when the regressors are functional random fields. More precisely, we consider $Z_i=(X_i,Y_i)$, $i{\in}\mathbb{N}^N$ be a $\mathcal{F}{\times}\mathbb{R}$-valued measurable strictly stationary spatial process, where $\mathcal{F}$ is a semi-metric space and we study the spatial interaction of $X_i$ and $Y_i$ via the robust estimation for the regression function. We propose a family of robust nonparametric estimators for regression function based on the kernel method. The main result of this work is the establishment of the asymptotic normality of these estimators, under some general mixing and small ball probability conditions.

• Communications for Statistical Applications and Methods
• /
• 제26권6호
• /
• pp.591-610
• /
• 2019

Deep Learning is one of the machine learning methods to find features from a huge data using non-linear transformation. It is now commonly used for supervised learning in many fields. In particular, Convolutional Neural Network (CNN) is the best technique for the image classification since 2012. For users who consider deep learning models for real-world applications, Keras is a popular API for neural networks written in Python and also can be used in R. We try examine the parameter estimation procedures of Deep Neural Network and structures of CNN models from basics to advanced techniques. We also try to figure out some crucial steps in CNN that can improve image classification performance in the CIFAR10 dataset using Keras. We found that several stacks of convolutional layers and batch normalization could improve prediction performance. We also compared image classification performances with other machine learning methods, including K-Nearest Neighbors (K-NN), Random Forest, and XGBoost, in both MNIST and CIFAR10 dataset.

• 한국환경과학회지
• /
• 제7권4호
• /
• pp.511-522
• /
• 1998

The two-stage numerical model was used to study the relation between three-dimensional local wind seal area for Korean peninsula. The first stave is three dimensional time-dependent local wind model which elves the wind field and vertical diffusion coefncient. The second stage is advection/duusion model which uses the results of the first stage as input data. First, wand fields on Korean peninsula for none synoptic scale wand showed typical land and sea breeze circulation, and the emitted particles were transported by sea breeze for daytime, emissions return to sea by land breeze for nighttime.

• 한국어정보학회:학술대회논문집
• /
• 한국어정보학회 2016년도 제28회 한글및한국어정보처리학술대회
• /
• pp.321-323
• /
• 2016

최근 연구에서 기계학습 중 지도학습 방법으로 개체명 인식을 하고 있다. 그러나 지도 학습 방법은 데이터를 만드는 비용과 시간이 많이 필요로 한다. 본 연구에서는 주석 된 말뭉치를 사용하여 지도 학습 방법을 사용 한다. 의생명 개체명 인식은 Protein, RNA, DNA, Cell type, Cell line 등을 포함한 텍스트 처리에 중요한 기초 작업입니다. 그리고 의생명 지식 검색에서 가장 기본과 핵심 작업 중 하나이다. 본 연구에서는 순환형 신경망과 워드 임베딩을 자질로 사용한 조건적 임의 필드에 대한 성능을 비교한다. 조건적 임의 필드에 N_Gram만을 자질로 사용한 것을 기준점으로 설정 하였고, 기준점의 결과는 70.09% F1 Score이다. RNN의 jordan type은 60.75% F1 Score, elman type은 58.80% F1 Score의 성능을 보여준다. 조건적 임의 필드에 CCA, GLOVE, WORD2VEC을 사용 한 결과는 각각 72.73% F1 Score, 72.74% F1 Score, 72.82% F1 Score의 성능을 얻을 수 있다.

• 품질경영학회지
• /
• 제6권1호
• /
• pp.23-26
• /
• 1978

The purpose of this study is to find the problems, which occurs in the course of the acceptance and the propulsion of the Q. C. circle, the accomodational attitude, and the desirable circumstances for the Q. C. circles, to be of help to settle firmly the "Sae-ma-eul" (New-Community) Movement in the industrial fields, and Q. C. circle and to fix the Q. C. of Korean style through the investigation of the 310 Q. C. circles, which was selected at random from entered circles to the Q. C. Propulsion Headquarters until July, 1977, through the questionaire and interviews.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1995년도 추계학술대회 논문집
• /
• pp.672-675
• /
• 1995

Because of a sudden growth of the research of fatigue failure, recent machines or structures have been designed by damage tolerance design in many fields. Consequently, it is the most primary factor to clarity the specific character of fatique failure in the design of machines or structures considering reliability. A statistical analysis is required to analyze the outcome of an experiment or a life estimate by reason of that fatigue failure contains lots of random elements. Reliability analysis which has tukenn the place of the existing analyses in the consideration of the uncertainty of a material, is a very efficient way. Even reliability analysis, however, is not a perfect way to analyses the uncertainties of all the materials. This thesis would refer to a newly conceived data analysis that the coefficient of a system could cause the ambiguity of the relationship of an input and output.

• 한국산업융합학회 논문집
• /
• 제23권3호
• /
• pp.515-521
• /
• 2020

Deep learning methods have been effectively used to provide great improvement in various research fields such as machine learning, image processing and computer vision. One of the most frequently used deep learning methods in image processing is the convolutional neural networks. Compared to the traditional artificial neural networks, convolutional neural networks do not use the predefined kernels, but instead they learn data specific kernels. This property makes them to be used as feature extractors as well. In this study, we compared the quality of CNN features for traditional texture feature extraction methods. Experimental results demonstrate the superiority of the CNN features. Additionally, the recognition process and result of a pioneering CNN on MNIST database are presented.

• International Journal of Aeronautical and Space Sciences
• /
• 제18권4호
• /
• pp.816-826
• /
• 2017

In this paper, a generalized discretization scheme is proposed that can derive general-order finite difference equations representing the joint probability density function of dynamic response of stochastic systems. The various order of finite difference equations are applied to solutions of the Fokker-Planck-Kolmogorov (FPK) equation. The finite difference equations derived by the proposed method can greatly increase accuracy even at the tail parts of the probability density function, giving accurate reliability estimations. Compared with exact solutions and finite element solutions, the generalized finite difference method showed increasing accuracy as the order increases. With the proposed method, it is allowed to use different orders and types (i.e. forward, central or backward) of discretization in the finite difference method to solve FPK and other partial differential equations in various engineering fields having requirements of accuracy or specific boundary conditions.