베이지안 혼합 정규 분포를 이용한 선박 재항시간 분포의 추정
Estimation of Distributions of the Ship’s Time in a Port through Bayesian Mixture Normal Distribution
0
조회
0
다운로드
3
인용
0
참고문헌
키워드
초록
This study adopted Bayesian mixture normal distribution to estimate distributions of the ship’s time in a port. Data on the time in a port for full-container vessels which arrived at Kwangyang port and Incheon port with loading/unloading purpose in 2009 were used for empirical analysis. In the case of Kwangyang port data and Incheon port data, five and four were decided as the best component numbers of mixture normal distribution. Groups with the short vessel’s time in a port showed a tendency that the time converged on the average of it, but groups with the long vessel’s time in a port had marked variation in the time. Comparison of information-based criteria and fitted values of estimation results of mixture normal distribution and Erlang distribution revealed that the former distribution was superior to the latter distribution with regard to the fitting performance. This study will contribute to increasing explanatory power of the model by describing the heterogeneity of the ship’s time in a port. Finally, analysis results can be used for building a queueing model and simulation model.
Abstract
This study adopted Bayesian mixture normal distribution to estimate distributions of the ship’s time in a port. Data on the time in a port for full-container vessels which arrived at Kwangyang port and Incheon port with loading/unloading purpose in 2009 were used for empirical analysis. In the case of Kwangyang port data and Incheon port data, five and four were decided as the best component numbers of mixture normal distribution. Groups with the short vessel’s time in a port showed a tendency that the time converged on the average of it, but groups with the long vessel’s time in a port had marked variation in the time. Comparison of information-based criteria and fitted values of estimation results of mixture normal distribution and Erlang distribution revealed that the former distribution was superior to the latter distribution with regard to the fitting performance. This study will contribute to increasing explanatory power of the model by describing the heterogeneity of the ship’s time in a port. Finally, analysis results can be used for building a queueing model and simulation model.
참고문헌 (20)
- 김달호, "R과 WINBUGS를 이용한 베이지안 통계학" (2005), 자유아카데미
- 김창곤, "항만 대기시스템에서 시간분포의 통계적 검증에 대한 사례연구", 해양정책연구, 12 (1997)
- 김형태, "항만하역장비 현대화자금 지원제도 도입 절실하다", 월간 해양수산, (247) (2005)
- 박병인, "이질적 복수서버를 갖는 혼잡 컨테이너터미널의 선박관련 시간분포 추정", 해양정책연구, 13(1) (1998)
- 백인흠, "선박재항시간에 대한 분석연구: 인천항의 경우", 수산해양교육연구, 10(1) (1998)
- 사공훈, "국내 벌크선박의 체선원인에 관한 실증적 분석 ― 접안대기시간을 중심으로 ―", 해운물류연구, 25(2), pp.263~284 (2009), 한국해운물류학회
- 신강원, "생존분석모형을 이용한 선박의 재항시간 및 온실가스 배출량 분석", 대한토목학회 논문집D, 30(4), pp.323~330 (2010), 대한토목학회
- 윤신휘, "기계학습 기법을 이용한 본선의 재항시간 예측", 석사, pp.2009
- 장영태, "우리나라 주요 수출입 항만에서의 선박 입출항시간 분포 추정에 관한 연구", 한국해운학회지, 19 (1994)
- Dempster, A. P., "Maximum Likelihood from Incomplete Data via the EM Algorithm", Journal of the Royal Statistical Society Series B, 39 (1977)
- Diebolt, J., "Estimation of Finite Mixture Distributions through Bayesian Sampling", Journal of the Royal Statistical Society Series B, 56 (1994)
- Frühwirth-Schnatter, S., "Finite Mixture and Markov Switching Models" (2006), Springer
- Lo, Y., "Testing the Number of Components in a Normal Mixture", Biometrika, 88 (2001)
- McLachlan, G., "Finite Mixture Model" (2000), JohnWiley & Sons
- Nylund, K. L., "Deciding on the Number of Classes in Latent Class Analysis and Growth Mixture Modeling: A Monte Carlo Simulation Study", Structural Equation Modeling, 14 (2007)
- Page, E., "Queueing Theory in OR" (1972), Butterworths
- Park, B. J., "Bayesian Mixture Modeling Approach to Account for Heterogeneity in Speed Data", Transportation Research Part B, 44 (2010)
- Rossi, P. E., "bayesm: Bayesian Inference for Marketing", Micro-econometrics, R package version 2, 2007, pp.1~3