• 한국지하수토양환경학회지:지하수토양환경
• /
• 제17권5호
• /
• pp.56-67
• /
• 2012

In the present study, an enhanced subsurface prediction algorithm based on a non-parametric geostatistical model and a history matching technique through Gibbs sampler is developed and the iterative prediction improvement procedure is proposed. The developed model is applied to a simple two-dimensional synthetic case where domain is composed of three different hydrogeologic media with $500m{\times}40m$ scale. In the application, it is assumed that there are 4 independent pumping tests performed at different vertical interval and the history curves are acquired through numerical modeling. With two hypothetical borehole information and pumping test data, the proposed prediction model is applied iteratively and continuous improvements of the predictions with reduced uncertainties of the media distribution are observed. From the results and the qualitative/quantitative analysis, it is concluded that the proposed model is good for the subsurface prediction improvements where the history data is available as a supportive information. Once the proposed model be a matured technique, it is believed that the model can be applied to many groundwater, geothermal, gas and oil problems with conventional fluid flow simulators. However, the overall development is still in its preliminary step and further considerations needs to be incorporated to be a viable and practical prediction technique including multi-dimensional verifications, global optimization, etc. which have not been resolved in the present study.

• 한국측량학회지
• /
• 제26권1호
• /
• pp.85-91
• /
• 2008

수평조정망과 같이 커다란 희소행렬(sparse matrix)을 계산할 때, 시간적 효율 및 공간적 효율을 높이기 위해서 재배열(reordering) 과정을 거치게 된다. 본 연구에서는 SMMS(Sparse Matrix Manip ulation System) 프로그램을 이용해서 희소행렬의 원소를 각각의 재배열 방법으로 재배열 한 후, 전체 계산에 걸리는 시간과 치환배열을 구해 해를 구하는 과정시 발생하는 Fill-in의 개수를 계산해서 각 방법의 효율성을 비교하였다. 그 결과, Minimum Bandwidth 기반의 GPS(Gibbs-Poole-Stockmeyer), RCM(Reverse Cuthill-Mckee) 방법보다 최소 차수(Minimum Degree) 기반의 MD(Minimum Degree), MMD(Mutiple Minimum Degree) 방법이 더 효율적인 모습을 보여주었다. 하지만, 행렬의 원소 분포에 따라서 최적의 성능을 보이는 재배열 방법은 달라질 수 있다는 것을 알 수 있었다. 이러한 연구 결과는 향후 전국 기준점의 좌표값 재조정 시, 또는 대규모 측지망 조정 등에서 구성 요소 계산에 필요한 시간, 저장 공간 등의 효율을 높일 수 있는 효과를 기대할 수 있을 것이라 사료된다.

• Communications for Statistical Applications and Methods
• /
• 제21권1호
• /
• pp.45-60
• /
• 2014

We introduce a MCMC sampling for a generalized linear normal random effects model with the ordered probit link function based on latent variables from suitable truncated normal distribution. Such models have proven useful in practice and we have observed numerically reasonable results in the estimation of fixed effects when the random effect term is provided. Applications that utilize Korean Welfare Panel Study data can be difficult to model; subsequently, we find that an ordered probit model with the random effects leads to an improved analyses with more accurate and precise inferences.

• Journal of the Korean Statistical Society
• /
• 제29권4호
• /
• pp.407-422
• /
• 2000

This article is concerned with the selecting predictor variables to be included in building a class of binary response t-link regression models where both probit and logistic regression models can e approximately taken as members of the class. It is based on a modification of the stochastic search variable selection method(SSVS), intended to propose and develop a Bayesian procedure that used probabilistic considerations for selecting promising subsets of predictor variables. The procedure reformulates the binary response t-link regression setup in a hierarchical truncated normal mixture model by introducing a set of hyperparameters that will be used to identify subset choices. In this setup, the most promising subset of predictors can be identified as that with highest posterior probability in the marginal posterior distribution of the hyperparameters. To highlight the merit of the procedure, an illustrative numerical example is given.

• Communications for Statistical Applications and Methods
• /
• 제6권1호
• /
• pp.221-235
• /
• 1999

Let be $X_1$,...,$X_p$, $p\geq2$ independent random variables where each $X_i$ has a gamma distribution with $\textit{k}_i$ and $\theta_i$ The problem is to simultaneously estimate $\textit{p}$ gamma parameters $\theta_i$ and $\theta_i{^-1}$ under entropy loss where the parameters are believed priori. Hierarch ical Bayes(HB) and empirical Bayes(EB) estimators are investigated. And a preference of HB estimator over EB estimator is shown using Gibbs sampler(Gelfand and Smith 1990). Finally computer simulation is studied to compute the risk percentage improvements of the HB estimator and the estimator of Dey Ghosh and Srinivasan(1987) compared to UMVUE estimator of $\theta^{-1}$.

• 한국정보통신학회논문지
• /
• 제6권8호
• /
• pp.1146-1152
• /
• 2002

This paper proposed a scheme for finding an optimal adaptive data association for multi-target between measurements and tracks. First, we assume the relationships between measurements as Mrkov Random Field. Also assumed a priori of the associations as a Gibbs distribution. Based on these assumptions, it was possible to reduce the MAP estimate of the association matrix to the energy minimization problem. After then, we defined an energy function over the measurement space, that may incorporate most of the important natural constraints. Through the experiments, we analyzed and compared this algorithm with other representative algorithms. The result is that it is stable, robust, fast enough for real timecomputation, as well as more accurate than other methods.

• 한국지구물리탐사학회:학술대회논문집
• /
• 한국지구물리탐사학회 2003년도 Proceedings of the international symposium on the fusion technology
• /
• pp.340-343
• /
• 2003

This study presents a practical procedure for the Bayesian inversion of geophysical data by Markov chain Monte Carlo (MCMC) sampling and geostatistics. We have applied geostatistical techniques for the acquisition of prior model information, and then the MCMC method was adopted to infer the characteristics of the marginal distributions of model parameters. For the Bayesian inversion of dipole-dipole array resistivity data, we have used the indicator kriging and simulation techniques to generate cumulative density functions from Schlumberger array resistivity data and well logging data, and obtained prior information by cokriging and simulations from covariogram models. The indicator approach makes it possible to incorporate non-parametric information into the probabilistic density function. We have also adopted the MCMC approach, based on Gibbs sampling, to examine the characteristics of a posteriori probability density function and the marginal distribution of each parameter. This approach provides an effective way to treat Bayesian inversion of geophysical data and reduce the non-uniqueness by incorporating various prior information.

• Communications for Statistical Applications and Methods
• /
• 제19권5호
• /
• pp.705-720
• /
• 2012

In this paper, we introduce a hierarchical Bayesian model to simultaneously estimate the thresholds of each 6 cities. It was noted in the literature there was a dramatic increases in the number of deaths if the mean temperature passes a certain value (that we call a threshold). We estimate the difference of mortality before and after the threshold. For the hierarchical Bayesian analysis, some proper prior distribution of parameters and hyper-parameters are assumed. By combining the Gibbs and Metropolis-Hastings algorithm, we constructed a Markov chain Monte Carlo algorithm and the posterior inference was based on the posterior sample. The analysis shows that the estimates of the threshold are located at $25^{\circ}C{\sim}29^{\circ}C$ and the mortality around the threshold changes from -1% to 2~13%.

• Journal of the Korean Statistical Society
• /
• 제31권1호
• /
• pp.63-75
• /
• 2002

Most attempts at Bayesian analysis of neural networks involve hierarchical modeling. We believe that similar results can be obtained with simpler models that require less computational effort, as long as appropriate restrictions are placed on parameters in order to ensure propriety of posterior distributions. In particular, we adopt a model first introduced by Lee (1999) that utilizes an improper prior for all parameters. Straightforward Gibbs sampling is possible, with the exception of the bias parameters, which are embedded in nonlinear sigmoidal functions. In addition to the problems posed by nonlinearity, direct sampling from the posterior distributions of the bias parameters is compounded due to the duplication of hidden nodes, which is a source of multimodality. In this regard, we focus on sampling from the marginal posterior distribution of the bias parameters with Markov chain Monte Carlo methods that combine traditional Metropolis sampling with a slice sampler described by Neal (1997, 2001). The methods are illustrated with data examples that are largely confined to the analysis of nonparametric regression models.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2010년도 학술발표회
• /
• pp.581-585
• /
• 2010

This study introduced a Bayesian based frequency analysis in which the statistical trend seasonal analysis for hydrologic extreme series is incorporated. The proposed model employed Gumbel and GEV extreme distribution to characterize extreme events and a fully coupled bayesian frequency model was finally utilized to estimate design rainfalls in Seoul. Posterior distributions of the model parameters in both trend and seasonal analysis were updated through Markov Chain Monte Carlo Simulation mainly utilizing Gibbs sampler. This study proposed a way to make use of nonstationary frequency model for dynamic risk analysis, and showed an increase of hydrologic risk with time varying probability density functions. In addition, full annual cycle of the design rainfall through seasonal model could be applied to annual control such as dam operation, flood control, irrigation water management, and so on. The proposed study showed advantage in assessing statistical significance of parameters associated with trend analysis through statistical inference utilizing derived posterior distributions.