Premium
Predictive distributions in risk analysis and estimation for the triangular distribution
Author(s) -
Joo Yongsung,
Casella George
Publication year - 2001
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.489
Subject(s) - log normal distribution , mathematics , statistics , quantile , monte carlo method , triangular distribution , econometrics , distribution (mathematics) , inverse distribution , estimation , distribution fitting , heavy tailed distribution , probability distribution , uniform distribution (continuous) , economics , mathematical analysis , management
Many Monte Carlo simulation studies have been done in the field of risk analysis. This article demonstrates the importance of using predictive distributions (the estimated distributions of the explanatory variable accounting for uncertainty in point estimation of parameters) in the simulations. We explore different types of predictive distributions for the normal distribution, the lognormal distribution and the triangular distribution. The triangular distribution poses particular problems, and we found that estimation using quantile least squares was preferable to maximum likelihood. Copyright © 2001 John Wiley & Sons, Ltd.