• 제목/요약/키워드: Bayes estimate

검색결과 80건 처리시간 0.02초

An Estimation of Loss Ratio Based on Empirical Bayes Credibility

  • Lee, Kang Sup;Lee, Hee Chun
    • Communications for Statistical Applications and Methods
    • /
    • 제9권2호
    • /
    • pp.381-388
    • /
    • 2002
  • It has been pointed out that the classical credibility model used in Korea since the beginning of 1990's lacks in objectiveness. Recently, in order to improve objectiveness, the empirical Bayes credibility model utilizing general exposure units like the number of claims and premium has been employed, but that model itself is not quite applicable in the country like Korea whose annual and classified empirical data are not well accumulated and even varied severely. In this article, we propose a new and better model, Based on the new model, we estimate both credibility and loss ratio of each class for fire insurance plans by Korean insurance companies. As a conclusion, we empirically make sure analysis that the number of claims is a more reasonable exposure unit than premium.

Maximum penalized likelihood estimation for a stress-strength reliability model using complete and incomplete data

  • Hassan, Marwa Khalil
    • Communications for Statistical Applications and Methods
    • /
    • 제25권4호
    • /
    • pp.355-371
    • /
    • 2018
  • The two parameter negative exponential distribution has many practical applications in queuing theory such as the service times of agents in system, the time it takes before your next telephone call, the time until a radioactive practical decays, the distance between mutations on a DNA strand, and the extreme values of annual snowfall or rainfall; consequently, has many applications in reliability systems. This paper considers an estimation problem of stress-strength model with two parameter negative parameter exponential distribution. We introduce a maximum penalized likelihood method, Bayes estimator using Lindley approximation to estimate stress-strength model and compare the proposed estimators with regular maximum likelihood estimator for complete data. We also introduce a maximum penalized likelihood method, Bayes estimator using a Markov chain Mote Carlo technique for incomplete data. A Monte Carlo simulation study is performed to compare stress-strength model estimates. Real data is used as a practical application of the proposed model.

음성인식을 위한 변환 공간 모델에 근거한 순차 적응기법 (Sequential Adaptation Algorithm Based on Transformation Space Model for Speech Recognition)

  • 김동국;장준혁;김남수
    • 음성과학
    • /
    • 제11권4호
    • /
    • pp.75-88
    • /
    • 2004
  • In this paper, we propose a new approach to sequential linear regression adaptation of continuous density hidden Markov models (CDHMMs) based on transformation space model (TSM). The proposed TSM which characterizes the a priori knowledge of the training speakers associated with maximum likelihood linear regression (MLLR) matrix parameters is effectively described in terms of the latent variable models. The TSM provides various sources of information such as the correlation information, the prior distribution, and the prior knowledge of the regression parameters that are very useful for rapid adaptation. The quasi-Bayes (QB) estimation algorithm is formulated to incrementally update the hyperparameters of the TSM and regression matrices simultaneously. Experimental results showed that the proposed TSM approach is better than that of the conventional quasi-Bayes linear regression (QBLR) algorithm for a small amount of adaptation data.

  • PDF

IMPLEMENTATION OF DATA ASSIMILATION METHODOLOGY FOR PHYSICAL MODEL UNCERTAINTY EVALUATION USING POST-CHF EXPERIMENTAL DATA

  • Heo, Jaeseok;Lee, Seung-Wook;Kim, Kyung Doo
    • Nuclear Engineering and Technology
    • /
    • 제46권5호
    • /
    • pp.619-632
    • /
    • 2014
  • The Best Estimate Plus Uncertainty (BEPU) method has been widely used to evaluate the uncertainty of a best-estimate thermal hydraulic system code against a figure of merit. This uncertainty is typically evaluated based on the physical model's uncertainties determined by expert judgment. This paper introduces the application of data assimilation methodology to determine the uncertainty bands of the physical models, e.g., the mean value and standard deviation of the parameters, based upon the statistical approach rather than expert judgment. Data assimilation suggests a mathematical methodology for the best estimate bias and the uncertainties of the physical models which optimize the system response following the calibration of model parameters and responses. The mathematical approaches include deterministic and probabilistic methods of data assimilation to solve both linear and nonlinear problems with the a posteriori distribution of parameters derived based on Bayes' theorem. The inverse problem was solved analytically to obtain the mean value and standard deviation of the parameters assuming Gaussian distributions for the parameters and responses, and a sampling method was utilized to illustrate the non-Gaussian a posteriori distributions of parameters. SPACE is used to demonstrate the data assimilation method by determining the bias and the uncertainty bands of the physical models employing Bennett's heated tube test data and Becker's post critical heat flux experimental data. Based on the results of the data assimilation process, the major sources of the modeling uncertainties were identified for further model development.

Bayes 풍의 RFID Tag 인식 (Bayesian Cognizance of RFID Tags)

  • 박진경;하준;최천원
    • 대한전자공학회논문지TC
    • /
    • 제46권5호
    • /
    • pp.70-77
    • /
    • 2009
  • 하나의 reader와 여러 tag로 구성된 RFID 망에서 tag의 응답 간 충돌을 중재하기 위해 tag가 응답하도록 여러 슬롯을 마련해 주는 프레임화 및 슬롯화된 ALOHA 방식이 소개되었다. 프레임화 및 슬롯화된 ALOHA에서는 tag 인식의 효율이 극대화되기 위해 프레임 별 슬롯의 수가 최적화되어야 한다. 이러한 최적화는 tag의 수를 필요로 하나 reader는 tag의 수를 알기 힘들다. 본 논문에서는 별도로 tag의 수를 추정하지 않고 슬롯의 수에 대해 직접 Bayes action을 취하는 프레임화 및 슬롯화된 ALOHA에 기초한 tag 인식 방식을 제안한다. 구체적으로 Bayes action은 tag의 수가 갖는 사전 분포, 어떤 tag도 응답하지 않은 슬롯의 수에 대한 관찰값, 그리고 인식률을 반영한 손실 함수를 결합한 결정 문제를 풀어 구한다. 또한 tag의 수가 갖는 사전 분포의 진화를 통해 각 프레임에서 이러한 Bayes action을 지원한다. 모의 실험 결과로부터 진화하는 사전 분포와 Bayes action의 쌍은 robust 방식을 이루어 tag의 수의 참값과 초기 추측값의 큰 괴리에도 불구하고 일정 수준의 인식률을 얻을 수 있음을 관찰한다. 또한 제안하는 방식은 tag의 수에 대한 고전적인 추정값을 사용하는 방식에 비해 높은 인식 완료 확률을 얻을 수 있음을 확인한다.

기하분포에 기초한 관리도에서 베이즈추정량과 최대우도추정량 사용의 성능 비교 (Comparisons of the Performance with Bayes Estimator and MLE for Control Charts Based on Geometric Distribution)

  • 홍휘주;이재헌
    • 응용통계연구
    • /
    • 제28권5호
    • /
    • pp.907-920
    • /
    • 2015
  • 기하분포에 기초한 관리도는 불량품이 드물게 발생하는 고품질공정에서 불량률의 변화를 효율적으로 탐지할 수 있다고 알려져 있다. 이러한 관리도를 사용할 때 기본적인 가정은 관리상태일 때의 불량률이 알려져 있거나 또는 정확하게 추정되었다는 것이다. 그러나 고품질공정에서 불량률은 아주 작기 때문에 이를 정확하게 추정하기가 쉽지 않으며 또한 아주 큰 표본크기가 필요한 경우도 종종 발생한다. 일반적으로 제1국면에서 관리상태의 불량률을 추정할 때 최대우도추정량을 사용하지만, 이 논문에서는 베이즈추정량의 사용을 제안하였다. 베이즈추정량을 사용할 경우 실무자의 사전지식을 반영할 수 있으며 표본에 불량품이 발견되지 않을 경우 발생하는 최대우도추정량의 문제점을 해결할 수 있다는 장점이 있다. 기하 관리도와 기하누적합 관리도에서 베이즈추정량을 사용한 경우와 최대우도추정량을 사용한 경우를 비교한 결과, 표본의 크기가 크지 않은 경우 베이즈추정량을 사용하는 것의 효율이 더 좋음을 알 수 있었다.

배열 안테나의 상관성 신호에서 원하는 신호 추정 방법에 대한 연구 (A Study on Desired Signal Estimation in Correlation Signal of Array Antennas)

  • 이민수
    • 한국정보전자통신기술학회논문지
    • /
    • 제8권4호
    • /
    • pp.275-280
    • /
    • 2015
  • 본 논문에서는 수정된 MUSIC 도래 방향 추정 알고리즘에 대해서 연구 하였다. 수정된 MUSIC 알고리즘은 특이 값 행렬과 베이즈 방법을 적용시켜 공 분산행렬을 최적화 시키고 가중치를 갱신하여 원하는 신호를 추정하는 방법이다. 그리고 MUSCI 알고리즘의 신호 부 공간 방법을 적용시켜 원하는 신호를 정확히 추정하였다. 무상관 신호가 수신 시스템에 입사하면 기존의 MUSIC알고리즘으로 원하는 신호를 추정 할 수 있다. 그러나 일반적으로 수신 시스템에는 상관성 신호가 입사하므로 기존의 MUSIC알고리즘으로 원하는 도래 방향 신호를 추정할 수 있는 능력이 현저히 떨어진다. 모의실험을 통해서 상관성 신호인 경우에 본 연구에서 제안된 MUSIC알고리즘과 기존의 MUSIC알고리즘의 성능을 비교 분석한다.

신뢰도 데이터 합성 program의 개발 (A Development on Reliability Data Integration Program)

  • 이광원;박문희;오신규;한정민
    • 한국안전학회지
    • /
    • 제18권4호
    • /
    • pp.164-168
    • /
    • 2003
  • Bayes theorem, suggested by the British Mathematician Bayes (18th century), enables the prior estimate of the probability of an event under the condition given by a specific This theorem has been frequently used to revise the failure probability of a component or system. 2-Stage Bayesian procedure was firstly published by Shultis et al. (1981) and Kaplan (1983), and was further developed based on the studies of Hora & Iman (1990) Papazpgolou et al., Porn(1993). For a small observed failure number (below 12), the estimated reliability of a system or component is not reliable. In the case in which the reliability data of the corresponding system or component can be found in a generic reliability reference book, however, a reliable estimation of the failure probability can be realized by using Bayes theorem, which jointly makes use of the observed data (specific data) and the data found in reference book (generic data).

Estimators for Parameters Included in Cold Standby Systems with Imperfect Switches

  • Al-Ruzaiza A. S.;Sarhan Ammar M.
    • International Journal of Reliability and Applications
    • /
    • 제6권2호
    • /
    • pp.65-78
    • /
    • 2005
  • In this paper we derive estimations of the parameters included in the distribution of the lifetime of k-out-of-m cold standby system with imperfect switches. Maximum likelihood and Bayes procedures are followed to get such estimations. Numerical studies, using Monte Carlo simulation method, are given in order to explain how we can utilize the theoretical results derived, and to compare the performance of the two different methods used. The criterion of comparisons is the mean squared errors associated with each estimate.

  • PDF

Classical and Bayesian methods of estimation for power Lindley distribution with application to waiting time data

  • Sharma, Vikas Kumar;Singh, Sanjay Kumar;Singh, Umesh
    • Communications for Statistical Applications and Methods
    • /
    • 제24권3호
    • /
    • pp.193-209
    • /
    • 2017
  • The power Lindley distribution with some of its properties is considered in this article. Maximum likelihood, least squares, maximum product spacings, and Bayes estimators are proposed to estimate all the unknown parameters of the power Lindley distribution. Lindley's approximation and Markov chain Monte Carlo techniques are utilized for Bayesian calculations since posterior distribution cannot be reduced to standard distribution. The performances of the proposed estimators are compared based on simulated samples. The waiting times of research articles to be accepted in statistical journals are fitted to the power Lindley distribution with other competing distributions. Chi-square statistic, Kolmogorov-Smirnov statistic, Akaike information criterion and Bayesian information criterion are used to access goodness-of-fit. It was found that the power Lindley distribution gives a better fit for the data than other distributions.