𝔖 Bobbio Scriptorium
✦   LIBER   ✦

On the computation of aggregate claims distributions: some new approximations

✍ Scribed by Yogendra P. Chaubey; Jose´ Garrido; Sonia Trudeau

Publisher
Elsevier Science
Year
1998
Tongue
English
Weight
638 KB
Volume
23
Category
Article
ISSN
0167-6687

No coin nor oath required. For personal study only.

✦ Synopsis

This paper proposes a new approximation to the aggregate claims distribution based on the inverse Gaussian (IG) distribution. It is compared to several other approximations in the literature. The IG approximation compares favorably to the well-known gamma approximation. We also propose an IG-gamma mixture that approximates the true distribution extremely accurately, in a large variety of situations.

📜 SIMILAR VOLUMES

On the distributions of two classes of c
On the distributions of two classes of correlated aggregate claims
✍ Rohana S. Ambagaspitiya 📂 Article 📅 1999 🏛 Elsevier Science 🌐 English ⚖ 66 KB

We consider two classes of correlated aggregate claims distributions, one with univariate claim counts and multivariate claim sizes and the other with multivariate claim counts and univariate claim sizes. We show that when the multivariate claim counts distribution takes a certain form the second cl

On the distribution of a sum of correlat
On the distribution of a sum of correlated aggregate claims
✍ Rohana S. Ambagaspitiya 📂 Article 📅 1998 🏛 Elsevier Science 🌐 English ⚖ 246 KB

Assume that a book of business is the union of disjoint classes of business and each of which has an aggregate distribution. The classes of business are correlated. We present necessary formulas to compute the aggregate distribution for the whole book when claim counts distribution takes a certain f

On some new properties of the beta distr
On some new properties of the beta distribution
✍ Włodzimierz Krysicki 📂 Article 📅 1999 🏛 Elsevier Science 🌐 English ⚖ 94 KB

The ÿrst aim of this paper is to show how to present a random variable with the beta distribution (of the ÿrst kind) as a ÿnite or inÿnite product of independent random variable's (r.v.)'s X k , where k ∈ [1; 2; : : : ; n] or k ∈ N. Such a presentation of an r.v. with the gamma distribution in the f