• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.32 no.3
• /
• pp.99-109
• /
• 2009

Negative binomial yield model for semiconductor manufacturing consists of two parameters which are the average number of defects per die and the clustering parameter. Estimating the clustering parameter is quite complex because the parameter has not clear closed form. In this paper, a Bayesian approach using Markov Chain Monte Carlo is proposed to estimate the clustering parameter. To find an appropriate estimation method for the clustering parameter, two typical estimators, the method of moments estimator and the maximum likelihood estimator, and the proposed Bayesian estimator are compared with respect to the mean absolute deviation between the real yield and the estimated yield. Experimental results show that both the proposed Bayesian estimator and the maximum likelihood estimator have excellent performance and the choice of method depends on the purpose of use.

• The Korean Journal of Applied Statistics
• /
• v.24 no.1
• /
• pp.177-183
• /
• 2011

Hidden Markov random eld(HMRF) is one of the most common model for image segmentation which is an important preprocessing in many imaging devices. The HMRF has unknown hyper-parameters on Markov random field to be estimated in segmenting testing images. However, in practice, due to computational complexity, it is often assumed to be a fixed constant. In this paper, we numerically show that the segmentation results very depending on the fixed hyper-parameter, and, if the parameter is misspecified, they further depend on the choice of the class-labelling algorithm. In contrast, the HMRF with estimated hyper-parameter provides consistent segmentation results regardless of the choice of class labelling and the estimation method. Thus, we recommend practitioners estimate the hyper-parameter even though it is computationally complex.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.31 no.4
• /
• pp.31-49
• /
• 1989

It is of the most urgent necessity to get hydrological time series of long duration for the establishment of rational design and operation criterion for the Agricultural hydraulic structures. This study was conducted to select best fitted frequency distribution for the monthly runoff and to simulate long series of generated flows by multi-season first order Markov model with comparison of statistical parameters which are derivated from observed and sy- nthetic flows in the five watersheds along Geum river basin. The results summarized through this study are as follows. 1. Both two parameter gamma and two parameter lognormal distribution were judged to be as good fitted distributions for monthly discharge by Kolmogorov-Smirnov method for goodness of fit test in all watersheds. 2. Statistical parameters were obtained from synthetic flows simulated by two parameter gamma distribution were closer to the results from observed flows than those of two para- meter lognormal distribution in all watersheds. 3. In general, fluctuation for the coefficient of variation based on two parameter gamma distribution was shown as more good agreement with the observed flow than that of two parameter lognormal distribution. Especially, coefficient of variation based on two parameter lognormal distribution was quite closer to that of observed flow during June and August in all years. 4. Monthly synthetic flows based on two parameter gamma distribution are considered to give more reasonably good results than those of two parameter lognormal distribution in the multi-season first order Markov model in all watersheds. 5. Synthetic monthly flows with 100 years for eack watershed were sjmulated by multi- season first order Markov model based on two parameter gamma distribution which is ack- nowledged to fit the actual distribution of monthly discharges of watersheds. Simulated sy- nthetic monthly flows may be considered to be contributed to the long series of discharges as an input data for the development of water resources. 6. It is to be desired that generation technique of synthetic flow in this study would be compared with other simulation techniques for the objective time series.

• Journal of the Korean Society of Systems Engineering
• /
• v.14 no.1
• /
• pp.44-51
• /
• 2018

The purpose of RAM analysis in weapon systems is to reduce life cycle costs, along with improving combat readiness by meeting RAM target value. We analyzed the sensitivity of the RAM analysis parameters to the use of the operating system by using the Markov Process based model (MPS, Markov Process Simulation) developed for RAM analysis. A Markov process-based RAM analysis model was developed to analyze the sensitivity of parameters (MTBF, MTTR and ALDT) to the utility of the 81mm mortar. The time required for the application to reach the steady state is about 15,000H, which is about 2 years, and the sensitivity of the parameter is highest for ALDT. In order to improve combat readiness, there is a need for continuous improvement in ALDT.

• ETRI Journal
• /
• v.5 no.1
• /
• pp.10-16
• /
• 1983

When analyzing a complex system with repair environments, it is necessary to calculate such parameters as availability and various kinds of failure time measures. These measures are defined and methods of calculating them using Markov process are presented. Analyzing the various system states, numerical values of the reliability measures can be obtained by calculating the state probabilities. And these techniques are widely applied to reliability prediction and also to maintenance strategy.

• Proceedings of the IEEK Conference
• /
• 2000.11c
• /
• pp.29-32
• /
• 2000

HMM(Hidden Markov Model)은 시간적인 정보를 토대로 하는 수학적인 방법으로서 문자인식에 많이 사용되어지고 있다. 그런데 HMM이 적용되고자 하는 문제에서 사용되어지는 상태 수와 HMM에서 사용되어지는 parameter들은 처음에 결정되는 값들에 의해서 상당히 많은 영향을 받게 된다. 따라서 한글의 특성을 이용한 HMM의 상태 수를 결정한 후 결정되어진 각각의 HMM parameter들을 Genetic Algorithm을 이용하였다. Genetic Algorithm은 매개변수 최적화 문제에 대하여 자연의 진화 원리를 마땅한 알고리즘으로 선택, 교배, 돌연변이 연산을 이용하여 최적의 개체를 구하게 된다. 여기서는 HMM에서의 Viterbi Algorithm을 적합도 검사에 사용하였다.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.273-273
• /
• 2000

We present a new scheme to increase the performance of edge-preserving image smoothing from the parameter tuning of a Markov random field (MRF) function. The method is based on automatic control of the image smoothing-strength in MRF model ing in which an introduced parameter function is based on control of enforcing power of a discontinuity-adaptive Markov function and edge magnitude resulted from discontinuities of image intensity. Without any binary decision for the edge magnitude, adaptive control of the enforcing power with the full edge magnitude could improve the performance of discontinuity-preserving image smoothing.

• 제어로봇시스템학회:학술대회논문집
• /
• 1998.10a
• /
• pp.289-294
• /
• 1998

The distributed-parameter structures expressed with the partial differential equations are considered as the infinite-dimensional dynamic system. For implementation of a controller in multivariate systems, it is necessary to derive the state-space reduced order model. By the eigensystem realization algorithm, we can yield tile subspace system with the Markov parameters derived from the measured frequency response function by the inverse discrete Fourier transformation. We also review the necessary conditions for the convergence of the approximation system and the error bounds in terms of the singular values of Markov-parameter matrices. To determine the natural frequencies and modal damping ratios, the modal coordinate transformation is applied to the realization system. The vibration test for a smart structure is performed to provide the records of frequency response functions used in the subspace system realization.

• Coupled systems mechanics
• /
• v.11 no.5
• /
• pp.411-438
• /
• 2022

In this paper, we present the parameter identification for inelastic and multi-scale problems. First, the theoretical background of several fundamental methods used in the upscaling process is reviewed. Several key definitions including random field, Bayesian theorem, Polynomial chaos expansion (PCE), and Gauss-Markov-Kalman filter are briefly summarized. An illustrative example is given to assimilate fracture energy in a simple inelastic problem with linear hardening and softening phases. Second, the parameter identification using the Gauss-Markov-Kalman filter is employed for a multi-scale problem to identify bulk and shear moduli and other material properties in a macro-scale with the data from a micro-scale as quantities of interest (QoI). The problem can also be viewed as upscaling homogenization.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.19 no.8
• /
• pp.1915-1926
• /
• 1995

The Fatigue characteristics of 8-harness satin woven CFRP composites with a circular hole are experimentally investigated under constant amplitude tension-tension loading. It is found in this study that the fatigue damage accumulation behavior is very random and history-independent, and the fatigue cumulative damage is linearly related with the mean number of cycles to a specified damage state. From these results, it is known that the fatigue characteristics of CFRP composites satisfy the basic assumptions of Markov chain theory and the parameter of Markov chain model can be determined only by mean and variance of fatigue lives. The predicted distribution of the fatigue cumulative damage using Markov chain model shows a good agreement with the test results. For the fatigue life distribution, Markov chain model makes similar accuracy to 2-parameter Weibull distribution function.