• 정보처리학회논문지D
• /
• 제10D권2호
• /
• pp.277-282
• /
• 2003

웹으로부터 유용한 정보를 얻기 위한 연구는 현재 많이 진행되고 있다. 본 논문에서는 특히 웹 로그 데이터의 희소성에 대한 문제 해결과 이를 통한 웹 사용자의 군집화 방안에 대하여 연구하였다. MCMC 방법의 베이지안 추론에 의한 결측치 대체 기법을 이용하여 웹 데이터의 희소성을 제거하였고, 주성분에 의한 산점도를 통하여 형상지도의 차원을 결정한 자기 조직화지도를 이용하여 웹 사용자의 군집화를 수행하였다. 제안 기법은 기존의 방법들에 비해 모형의 정확도와 빠른 학습 시간을 제공하여 주었다. KDD Cup 데이터를 이용한 실험을 통하여 제안 방법에 대한 문제 해결 절차 및 성능 평가를 객관적으로 확인하였다.

• 한국전산구조공학회논문집
• /
• 제23권6호
• /
• pp.641-649
• /
• 2010

Estimation of uncertain parameters is required in many engineering problems which involve probabilistic structural analysis as well as prognosis of existing structures. In this case, Bayesian framework is often employed, which is to represent the uncertainty of parameters in terms of probability distributions conditional on the provided data. The resulting form of distribution, however, is not amenable to the practical application due to its complex nature making the standard probability functions useless. In this study, Markov chain Monte Carlo (MCMC) method is proposed to overcome this difficulty, which is a modern computational technique for the efficient and straightforward estimation of parameters. Three case studies that implement the estimation are presented to illustrate the concept. The first one is an inverse estimation, in which the unknown input parameters are inversely estimated based on a finite number of measured response data. The next one is a metamodel uncertainty problem that arises when the original response function is approximated by a metamodel using a finite set of response values. The last one is a prognostics problem, in which the unknown parameters of the degradation model are estimated based on the monitored data.

• Journal of the Korean Statistical Society
• /
• 제31권1호
• /
• pp.77-91
• /
• 2002

This paper considers a functional regression model with truncated errors in explanatory variables. We show that the ordinary least squares (OLS) estimators produce bias in regression parameter estimates under misspecified models with ignored errors in the explanatory variable measurements, and then propose methods for analyzing the functional model. Fully parametric frequentist approaches for analyzing the model are intractable and thus Bayesian methods are pursued using a Markov chain Monte Carlo (MCMC) sampling based approach. Necessary theories involved in modeling and computation are provided. Finally, a simulation study is given to illustrate and examine the proposed methods.

• Journal of the Korean Data and Information Science Society
• /
• 제19권4호
• /
• pp.1465-1476
• /
• 2008

In recent years wavelet methods have been focused on block shrinkage or thresholding approaches to accounting for the sparseness of the wavelet representation for an unknown function. The block shrinkage or thresholding methods have been developed in both of classical methods and Bayesian methods. In this paper, we propose a Bayesian approach to selecting wavelet basis functions at each resolution level without MCMC procedure. Simulation study and an application are shown.

• 한국신뢰성학회지:신뢰성응용연구
• /
• 제17권3호
• /
• pp.213-223
• /
• 2017

Purpose: The purpose of this study is to arrange the life data analysis literatures based on the Bayesian method quantitatively and provide it as tables. Methods: The Bayesian method produces a more accurate estimates of other traditional methods in a small sample size, and it requires specific algorithm and prior information. Based on these three characteristics of the Bayesian method, the criteria for classifying the literature were taken into account. Results: In many studies, there are comparisons of estimation methods for the Bayesian method and maximum likelihood estimation (MLE), and sample size was greater than 10 and not more than 25. In probability distributions, a variety of distributions were found in addition to the distributions of Weibull commonly used in life data analysis, and MCMC and Lindley's Approximation were used evenly. Finally, Gamma, Uniform, Jeffrey and extension of Jeffrey distributions were evenly used as prior information. Conclusion: To verify the characteristics of the Bayesian method which are more superior to other methods in a smaller sample size, studies in less than 10 samples should be carried out. Also, comparative study is required by various distributions, thereby providing guidelines necessary.

• Communications for Statistical Applications and Methods
• /
• 제11권2호
• /
• pp.381-397
• /
• 2004

The marginal likelihood has become an important tool for model selection in Bayesian analysis because it can be used to rank the models. We discuss the marginal likelihood for Poisson regression models that are potentially useful in small area estimation. Computation in these models is intensive and it requires an implementation of Markov chain Monte Carlo (MCMC) methods. Using importance sampling and multivariate density estimation, we demonstrate a computation of the marginal likelihood through an output analysis from an MCMC sampler.

• Smart Structures and Systems
• /
• 제15권3호
• /
• pp.735-749
• /
• 2015

The major difficulty of using Bayesian probabilistic inference for system identification is to obtain the posterior probability density of parameters conditioned by the measured response. The posterior density of structural parameters indicates how plausible each model is when considering the uncertainty of prediction errors. The Markov chain Monte Carlo (MCMC) method is a widespread medium for posterior inference but its convergence is often slow. The differential evolution adaptive Metropolis-Hasting (DREAM) algorithm boasts a population-based mechanism, which nms multiple different Markov chains simultaneously, and a global optimum exploration ability. This paper proposes an improved differential evolution adaptive Metropolis-Hasting algorithm (IDREAM) strategy to estimate the posterior density of structural parameters. The main benefit of IDREAM is its efficient MCMC simulation through its use of the adaptive Metropolis (AM) method with a mutation strategy for ensuring quick convergence and robust solutions. Its effectiveness was demonstrated in simulations on identifying the structural parameters with limited output data and noise polluted measurements.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2000년도 가을 학술발표논문집 Vol.27 No.2 (2)
• /
• pp.90-92
• /
• 2000

데이터마이닝 문제는 데이터를 그 속성들에 따라 분류하여 예측하는 것뿐만 아니라 분류된 속성들간의 연관성에 대해 잘 설명할 수 있어야 한다. 일반적으로 변수들간의 연관성을 잘 설명할 수 있으면서도 높은 예측력을 가지는 방법으로는 베이지안 네트웍 분류자(Bayesian network classifier)가 있다. 그러나 이것은 데이터 마이닝과 같은 대용량 데이터에서는 성능이 떨어지는 단점이 있다. 이에 이 논문에서는 최근 RBF 신경망이 입력변수 선정문제에 성공적으로 적용된 Reversible Jump Markov Chain Monte Carlo 방법을 이용하여 최적의 입력변수들만을 선택하여 베이지안 네트웍을 학습하는 Selective BN Augmented Naive-Bayes Classifier를 새로운 방안으로 제안하고 이를 실제 데이터마이닝 문제에 적용한 결과를 제시한다.

• Asian Journal of Innovation and Policy
• /
• 제8권1호
• /
• pp.31-57
• /
• 2019

Myanmar's current power situation remains severely constrained despite being richly endowed in primary energy sources. With low levels of electrification, the demand for power is not adequately met. Cooperation in energy has been a major focus of future initiative for all developed and developing nations. If we want to solve climate change, and change our energy infrastructure, we need to be innovative and entrepreneurial in energy generation. This paper will help us in examining Bayesian MCMC Analysis for the parameters estimation among the arrival rates of disaster occurrences, firm's expected income-based electricity tariffs, and estimated R&D investment expenses in new energy industry. Focusing on Japan's electric power business, we would like to search the potential for innovative initiatives in new technological energy industry for the regional development and ecological sustainability in Myanmar.

• Asian Journal of Innovation and Policy
• /
• 제7권3호
• /
• pp.625-645
• /
• 2018

This paper's objective is to draw a decision guideline to continue research and development (R&D) investments in biotech start-ups facing the "Valley of Death" syndrome - a long negative profit period during a financial crisis. The data include financial indices as Net income, Revenues, Total stockholders' equity, Cash & equivalents, and R&D expenses of 18 major biotech companies (nine in negative profit and nine positive, in FY2008) and 15 major pharmaceutical corporations as benchmarks both in FY2008 and in FY2016 derived from the US SEC Database, EDGAR. A first methodology dealing with real options analysis assumes Total stockholders' equity as a growth option. And a second methodology, Bayesian Markov chain Monte Carlo (MCMC) analysis, is applied to test the probability relationship between the Total stockholders' equity and the R&D expenses in these three groups. This study confirms that Total stockholders' equity can play the role of a call option to support continuing R&D investments even in negative profits.