Polynomial power-Pareto quantile function models
- 116 Downloads
In this paper we propose a polynomial power-Pareto quantile function model and a Bayesian method for parameters estimation. We also carried out simulation studies and applied our methodology to real data sets empirically. The results show that a quantile function approach to statistical modelling is very flexible due to the properties of quantile functions, and that the combination of a power and a Pareto distribution enables us to model both the main body and the tails of a distribution, even though the mathematical form of the distribution does not exist. Our research also suggests a new approach to studying extreme values based on a whole data set rather than group maximum/minimum or exceedances above/below a proper threshold value.
KeywordsBayesian methods Quantile functions Power-Pareto distribution Simulation Speed and stopping distance data Wave and surge data
AMS 2000 Subject Classifications62XX 62Fxx 62F15
Unable to display preview. Download preview PDF.
- Coles, S.: An Introduction to Statistical Modelling of Extreme Values. Springer Series in Statistics. Springer, New York (2001)Google Scholar
- Freeze, K.: Monster waves threaton rescu helicopters. In: Proceedings, U.S. Naval Institute, Annapolis, Maryland, p. 410. http://www.check-six.com/Coast_Guard/Monster_Waves_Reprint-screen.pdf (2006)
- Gilks, W.R., Richardson, S., Spiegelhalter, D.J.: Markov Chain Monte Carlo in Practice: Interdisciplinary Statistics. Chapman and Hall/CRC, London (1996)Google Scholar