333Kb
by PREDA,
Vasile and
CIUMARA,
Roxana
Published in Romanian Journal of Economic Forecasting, volume 7 issue 2,
2006.
|
| Requires a PDF viewer such as Xpdf
or Adobe
Acrobat Reader 333Kb |
In this paper we make a comparison between two composite models: lognormal-Pareto and Weibull-Pareto. The first one was introduced by Cooray and Ananda in 2005. The second composite distribution was constructed in the same manner as lognormal-Pareto. Here, we prove that these models behave similarly and they could be used in insurance bussiness for modelling actuarial data, especially in the cases where one deals with large loss payments.
Keywords:
composite models; lognormal, Weibull and Pareto distributions; maximum
likelihood estimation; smooth empirical estimation of percentils.
JEL Classification:
C16