• Communications for Statistical Applications and Methods
• /
• 제11권1호
• /
• pp.79-91
• /
• 2004

In this thesis, Bayesian parameter estimation procedure is discussed for the mean change model of multivariate normal random variates under the assumption of noninformative priors for all the parameters. Parameters are estimated by Gibbs sampling method. In Gibbs sampler, the change point parameter is generated by Metropolis-Hastings algorithm. We apply our methodology to numerical data to examine it.

• Journal of the Korean Data and Information Science Society
• /
• 제18권2호
• /
• pp.535-541
• /
• 2007

The Non-homogeneous Poisson process is probably the most popular model since it can model systems that are deteriorating or improving. The renewal process is a model that is often used to describe the random occurrence of events in time. But both these models are based on too restrictive assumptions on the effect of the repair action. The Modulated Power Law Process is a suitable model for describing the failure pattern of repairable systems when both renewal-type behavior and time trend are present. In this paper we propose maximum likelihood estimation of the next failure time after the system has experienced some failures, that is, Mean Time Between Failure for the MPLP model.

• 품질경영학회지
• /
• 제34권1호
• /
• pp.73-77
• /
• 2006

In this paper, the Bayesian procedure for the multiple test of fraction nonconforming, p, is proposed. It is the procedure for checking whether the process is out of control, in control, or under the permissible level for p. The procedure is as follows: first, setting up three types of models, $M_1:p=p_0,\;M_2:pp_0$, second, computing the posterior probability of each model. and then choosing the model with the largest posterior probability as a model most fitted for the observed sample among three competitive models. Finally, the simulation study is performed to examine the proposed method.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2005년도 추계학술발표대회 및 정기총회
• /
• pp.3-6
• /
• 2005

현재의 교육에 있어서 정보의 가치는 무한한 가능성을 내포하고 있다. 하지만 학습 과정에서 생성되는 많은 정보들이 재활용되지 못하고 있다. 학습 과정에서 학습자들이 생성한 결과물을 차기 학습에 참고자료로서 재활용될 경우 보다 체계적이고 질적으로 향상된 학습 과정 개발이 가능하며 또한 풍부한 실질적인 학습 참고 자료를 제공함으로써 학습 효과를 향상시킬 수 있다. 학습 과정에서 생성된 정보들을 체계적으로 관리, 검색 및 활용할 수 있는 교육용 저장소 관리 시스템이 요구된다. 이 논문에서는 이를 위한 E-Repository 시스템을 제안한다. E-Repository 시스템은 생성된 학습 결과물은 체계적으로 등록, 관리하고 학습들이 활용할 수 있는 기능을 제공함으로써 학습 효과를 배가시키고 보다 풍부한 정보의 제공 및 공유를 가능하게 한다. 이는 더 나아가 보다 체계적인 학습 과정 개발을 가능하게 할 것이다.

• Communications for Statistical Applications and Methods
• /
• 제2권2호
• /
• pp.231-242
• /
• 1995

In this paper we consider kernel type estimator of the spectral density at a point in the analysis of stationary time series data. The kernel entails choice of smoothing parameter called bandwidth. A data-based bandwidth choice is proposed, and it is obtained by solving an equation similar to Sheather(1986) which relates to the probability density estimation. A Monte Carlo study is done. It reveals that the spectral density estimates using the data-based bandwidths show comparatively good performance.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2006년도 추계 학술발표회 논문집
• /
• pp.167-180
• /
• 2006

The statistical survey and analysis is much restricted to time, space and material. The statistical survey and analysis could hardly resume. The statistical survey and analysis is very important to create various and accurate information. The statistical survey and analysis which is not a expert knowledge have many problems in productivity of information, reliability and etc. In this paper, we study the design and Implementation of web-based statistical survey and analysis consulting system which a client meet easily a statistical expert on the web.

• Journal of applied mathematics & informatics
• /
• 제9권1호
• /
• pp.289-301
• /
• 2002

This paper considers the Bayesian analysis of the regression model wish autoregressive errors. The Bayesian approach for finding the order p of autoregressive error is proposed and the proposed method can be simplified by generalized Savage-Dicky density ratio(Verdinelli and Wasser-man, [18]). And the Markov chain Monte Carlo method(Gibbs sample, [7]) is used in order to overcome the difficulty of Bayesian computations. Final1y, several examples are used to illustrate our proposed methodology.

• 한국품질경영학회:학술대회논문집
• /
• 한국품질경영학회 2006년도 춘계학술대회
• /
• pp.325-329
• /
• 2006

In this paper, the Bayesian procedure for the multiple test of fraction nonconforming, p, is proposed. It is the procedure for checking whether the process is out of control, in control, or under the permissible level for p. The procedure is as follows: first, setting up three types of models, $M_1:p=p_0,\;M_2:pp_0$, second, computing the posterior probability of each model. and then choosing the model with the largest posterior probability as a model most fitted for the observed sample among three competitive models. Finally, the simulation study is performed to examine the proposed method.

• 한국품질경영학회:학술대회논문집
• /
• 한국품질경영학회 2006년도 춘계학술대회
• /
• pp.319-324
• /
• 2006

In this paper, we propose Bayesian procedure for the multiple change points analysis in a sequence of fractions nonconforming. We first compute the Bayes factor for detecting the existence of no change, a single change or multiple changes. The Gibbs sampler with the Metropolis-Hastings subchain is run to estimate parameters of the change point model, once the number of change points is identified. Finally, we apply the results developed in this paper to both a real and simulated data.

• Communications for Statistical Applications and Methods
• /
• 제30권3호
• /
• pp.343-354
• /
• 2023

Recently, there have been many improvements in general language models using architectures such as GPT-3 proposed by Brown et al. (2020). Nevertheless, training complex models can hardly be done if the number of data is very small. Data augmentation that addressed this problem was more than normal success in image data. Image augmentation technology significantly improves model performance without any additional data or architectural changes (Perez and Wang, 2017). However, applying this technique to textual data has many challenges because the noise to be added is veiled. Thus, we have developed a novel method for performing data augmentation on text data. We divide the data into signals with positive or negative meaning and noise without them, and then perform data augmentation using k-doc augmentation to randomly combine signals and noises from all data to generate new data.