http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
약물독성시험에서 실험설계가 MTD의 결정에 미치는 영향
이윤동 ( Yoon Dong Lee ),이은경 ( Eun Kyung Lee ) 한국품질경영학회 2011 품질경영학회지 Vol.39 No.2
The purpose of Phase I clinical trial is to identify the maximum tolerated dose with specific toxicity rate. The standard TER design does not guarantee the pre-specified toxicity rate. It depends on the dose-toxicity curves. Therefore it is necessary to check the expected toxicity rate of various dose-toxicity curves before we conduct clinical trials. We developed TERAplusB library to help this situation, especially in cancer research. This package will help design the cancer clinical trial. We can compare the expected toxicity rates, the expected number of patients, and the expected times calculated with various dose-toxicity curves. This process will help find the best clinical trial design of the proposed drug.
이윤동(Yoon-Dong Lee),이은경(Eun-kyung Lee) 한국경영과학회 2013 韓國經營科學會誌 Vol.38 No.1
Diffusion is a basic mathematical tool for modern financial engineering. The theory of the estimation methods for diffusion models is an important topic of the financial engineering. Many researches have been tiled to apply the likelihood estimation method for estimating diffusion models. However, the likelihood estimation method for diffusion is complicated and needs much amount of computing. In this paper we develop the estimation methods which are simple enough to be compared to the Euler approximation method, and efficient enough statistically to be compared to the likelihood estimation method. We devise pseudo-likelihood and propose the maximum pseudo-likelihood estimation methods. The pseudo-likelihoods are obtained by approximating the transition density with normal distributions. The means and the variances of the distributions are obtained from the delta expansion suggested by Lee, Song and Lee (2012). We compare the newly suggested estimators with other existing estimators by simulation study. From the simulation study we find the maximum pseudo-likelihood estimator has very similar properties with the maximum likelihood estimator. Also the maximum pseudo-likelihood estimator is easy to apply to general diffusion models, and can be obtained by simple numerical steps.
선거 개표방송에서 출구조사 자료를 활용한 중간 득표율 추정에 관한 연구
이윤동 ( Yoon Dong Lee ),박진우 ( Jin Woo Park ) 한국조사연구학회 2011 조사연구 Vol.12 No.1
During major elections, three terrestrial broadcasting stations in Korea have covered the progresses of election results by announcing the simple sum of ballot counts of all ballot counting stations. The current approach, however, does not properly reflect the actual ballot count differences by ballot counting location, leading to cause unnecessary but possible confusions. In addition, the current coverage approach restricts the broadcasters from using regional poll data gained through exit polls by letting them to use the significant information on a one-off purpose to announce the initial prediction of the poll results and to fully disregard the exit poll results during the ballot counting process. Based on the understanding, this paper is designed to suggest a Bayesian approach to consolidate the exit poll results with the progressive ballot counting results and announce them as such. The suggested consolidation approach is expected to mitigate or avoid the possible confusions that may arise in connection with the different ballot counting paces by ballot counting station.
이윤동(Yoon Dong Lee) 건국대학교 경제경영연구소 2007 商經硏究 Vol.32 No.1
In academic area, R rapidly widen its territory, and rose as one of major statistical software. In this research, I review the general facts and methods for popularizing R at domestic level. I also consider the functions what is need to be implemented on R. I mentioned the necessity and importance of localizing the functions of R and translating help manual.
이윤동(Yoon Dong Lee) 건국대학교 경제경영연구소 2005 商經硏究 Vol.30 No.1
In this research, we review the penalized likelihood method, in the view of applying the method to spatial data analysis. We considered a few groups of specific topics. One is the topics of determining the weighting factor of penalty term. Determining the weighting factor is equivalent to determining the degree of freedom of the selected statistical model. In spatial statistics, infill asymptotic setting is a more reasonable framework than increasing domain asymptotic setting, but not a few research have considered the infill asymptotic setting of the weighting factor. We review the meaning of the topic. The other topic is to review the effects of method using data-adjusted bases. Differently with Kriging in spatial statistics and well known methods in time series models using data-adjusted bases for prediction, reproducing kernel Hilbert space method and fast Fourier transform method adopt the bases obtained independently from data. We considered the necessity of the comparative researches between the methods using data-adjusted bases and simple mathematical bases.
김대균,이윤동,Kim, Dae-Gyun,Lee, Yoon-Dong 한국통계학회 2011 응용통계연구 Vol.24 No.6
Since the research of Black and Scholes (1973), modeling methods using diffusion processes have performed principal roles in financial engineering. In modern financial theories, various types of diffusion processes were suggested and applied in real situations. An estimation of the model parameters is an indispensible step to analyze financial data using diffusion process models. Many estimation methods were suggested and their properties were investigated. This paper reviews the statistical properties of the, Euler approximation method, New Local Linearization(NLL) method, and Generalized Methods of Moment(GMM) that are known as the most practical methods. From the simulation study, we found the NLL and Euler methods performed better than GMM. GMM is frequently used to estimate the parameters because of its simplicity; however this paper shows the performance of GMM is poorer than the Euler approximation method or the NLL method that are even simpler than GMM. This paper shows the performance of the GMM is extremely poor especially when the parameters in diffusion coefficient are to be estimated. 현대금융공학에 있어서 확산모형은 중요한 역할을 담당하고 있다. 다양한 형태의 확산모형이 제안되어왔고 현실에 응용되어 왔다. 확산모형을 이용하여 금융자료를 분석하기 위하여는 확산모형의 모수를 추정하는 것이 필수불가결한 단계이다. 이들 모수에 대한 다양한 추정방법들이 제안되어 왔고, 많은 연구에서 이러한 추정방법들이 갖는 성질에 대하여 연구되어져왔다. 이 연구에서는 그 적용방법이 단순하여 가장 자주 사용되는 것으로 알려진, 오일러 근사법과 신국소근사법(NLL) 그리고 일반화 적률법(GMM)과 같은 세 가지 추정방법들에 대한 통계적 성질을 검토하게 될 것이다. 모의실험연구를 통하여 오일러근사법이나 NLL방법이 GMM 방법에 비하여 훨씬 좋은 성질을 가지고 있음을 보이게 된다. 특히 GMM은 적용방법이 단순할 뿐만 아니라 강건성(robustness)이라는 좋은 성질을 가지고 있는 것으로 알려져 있어서 많은 연구에서 매우 자주 사용되는 추정방법이다. 그러나 본 연구에서 확인해 본 바와 같이 GMM은 그 사용법이 오히려 더욱 단순한 NLL이나 오일러방법에 비하여 열등한 통계적 성질을 보여주고 있었다. 특히나 확산계수에 추정모수가 포함된 경우에 GMM은 매우 좋지 못한 성질을 보이게 된다.