• Journal of the Korean Statistical Society
• /
• 제29권1호
• /
• pp.17-27
• /
• 2000

In this paper, we develop noninformative priors that are used for estimating the reliability of stress-strength system under the Burr-type X model. A class of priors is found by matching the coverage probabilities of one-sided Bayesian credible interval with the corresponding frequentist coverage probabilities. It turns out that the reference prior as well as the Jeffreys prior are the second order matching prior. The propriety of posterior under the noninformative priors is proved. The frequentist coverage probabilities are investigated for samll samples via simulation study.

• Journal of the Korean Data and Information Science Society
• /
• 제13권2호
• /
• pp.235-242
• /
• 2002

In this paper, we develop noninformative priors that are used for estimating the ratio of failure rates under Freund's bivariate exponential distribution. A class of priors is found by matching the coverage probabilities of one-sided Baysian credible interval with the corresponding frequentist coverage probabilities. Also the propriety of posterior under the noninformative priors is proved and the frequentist coverage probabilities are investigated for small samples via simulation study.

• 한국산업정보학회논문지
• /
• 제4권4호
• /
• pp.47-52
• /
• 1999

본 논문에서는 Burr Type-X 분포하에서 stress-strength 의 신뢰도를 추정하는 데 사용되어지는 비정보적 사전분포들을 개발하는 것이다. 개발된 reference 사전분포가 first order matching 사전분포가 된다는 것이 밝혀졌으며, 또한 matching 사전분포하에서 사후분포의 타당성을 밝혔다. 소표본하에서, 고전적 포함확률들이 주여져 있다.

• Journal of the Korean Statistical Society
• /
• 제27권4호
• /
• pp.421-433
• /
• 1998

In this paper, matching priors for P(X < Y) are investigated when both distributions are exponential distributions. Two recent approaches for finding noninformative priors are introduced. The first one is the verger and Bernardo's forward and backward reference priors that maximizes the expected Kullback-Liebler Divergence between posterior and prior density. The second one is the matching prior identified by matching the one sided posterior credible interval with the frequentist's desired confidence level. The general forms of the second- order matching prior are presented so that the one sided posterior credible intervals agree with the frequentist's desired confidence levels up to O(n$^{-1}$ ). The frequentist coverage probabilities of confidence sets based on several noninformative priors are compared for small sample sizes via the Monte-Carlo simulation.

• Journal of the Korean Data and Information Science Society
• /
• 제21권3호
• /
• pp.569-574
• /
• 2010

We consider two components system which have Freund's bivariate exponential model. In this case, Bayesian multiple comparisons procedure for failure rates is sug-gested in K Freund's bivariate exponential populations. Here we assume that the com-ponents enter the study at random over time and the analysis is carried out at some prespeci ed time. We derive fractional Bayes factor for all comparisons under non- informative priors for the parameters and calculate the posterior probabilities for all hypotheses. And we select a hypotheses which has the highest posterior probability as best model. Finally, we give a numerical examples to illustrate our procedure.

• 응용통계연구
• /
• 제16권1호
• /
• pp.141-150
• /
• 2003

본 논문은 베이즈인자(Bayes factor)를 이용하여 정상(stationary) AR(1)모형의 자기회귀계수에 대해 다중검정하는 방법을 제시한다. 모수들에 대한 사전분포로는 무정보 사전분포(noninformative prior distribution)를 가정한다. 이러한 경우에 통상적으로 사용되는 베이즈인자를 근사없이 정확히 계산하여 각 모형에 대한 사후확률(posterior probability)을 얻는다. 최종적으로 모의실험 자료 및 실제 자료에 적용하여 이론의 결과가 잘 부합되는지를 검토한다.

• Journal of the Korean Data and Information Science Society
• /
• 제19권4호
• /
• pp.1391-1396
• /
• 2008

In this paper, we consider two components system which the lifetimes have Marshall and Olkin's bivariate exponential model with bivariate type I censored data. We propose a Bayesian independent test procedure for above model using fractional Bayes factor method by O'Hagan based on improper prior distributions. And we compute the fractional Bayes factor and the posterior probabilities for the hypotheses, respectively. Also we select a hypothesis which has the largest posterior probability. Finally a numerical example is given to illustrate our Bayesian testing procedure.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2005년도 공동추계학술대회
• /
• pp.509-515
• /
• 2005

Recently, support vector machines (SVMs) are being recognized as competitive tools as compared with other data mining techniques for solving pattern recognition or classification decision problems. Furthermore, many researches, in particular, have proved it more powerful than traditional artificial neural networks (ANNs)(Amendolia et al., 2003; Huang et al., 2004, Huang et al., 2005; Tay and Cao, 2001; Min and Lee, 2005; Shin et al, 2005; Kim, 2003). The classification decision, such as a binary or multi-class decision problem, used by any classifier, i.e. data mining techniques is cost-sensitive. Therefore, it is necessary to convert the output of the classifier into well-calibrated posterior probabilities. However, SVMs basically do not provide such probabilities. So it required to use any method to create probabilities (Platt, 1999; Drish, 2001). This study applies a method to estimate the probability of outputs of SVM to bankruptcy prediction and then suggests credit scoring methods using the estimated probability for bank's loan decision making.

• Journal of the Korean Data and Information Science Society
• /
• 제25권1호
• /
• pp.227-235
• /
• 2014

In this paper, we develop the noninformative priors for the scale parameter and the shape parameter in the log-logistic distribution. We developed the first and second order matching priors. It turns out that the second order matching prior matches the alternative coverage probabilities, and is a highest posterior density matching prior. Also we revealed that the derived reference prior is the second order matching prior for both parameters, but Jerffrey's prior is not a second order matching prior. We showed that the proposed reference prior matches the target coverage probabilities in a frequentist sense through simulation study, and an example based on real data is given.

• Communications for Statistical Applications and Methods
• /
• 제22권6호
• /
• pp.625-637
• /
• 2015

We develop a Bayesian clustering procedure based on a Dirichlet process prior with cluster specific random effects. Gibbs sampling of a normal mixture of linear mixed regressions with a Dirichlet process was implemented to calculate posterior probabilities when the number of clusters was unknown. Our approach (unlike its counterparts) provides simultaneous partitioning and parameter estimation with the computation of the classification probabilities. A Monte Carlo study of curve estimation results showed that the model was useful for function estimation. We find that the proposed Dirichlet process mixture model with cluster specific random effects detects clusters sensitively by combining vague edges into different clusters. Examples are given to show how these models perform on real data.