The rate of convergence for subexponential distributions
Abstract:
A distribution function F on the nonnegative real line is called subexponential if lims<inf>→∞</inf> (1 - F*<sup>n</sup>(x)}/(1 - F(x)) = n, for all n > 2, where F*<sup>n</sup> denotes the n-fold Stieltjes convolution of F with itself. In this paper, we consider the rate of convergence in the above definition and we discuss the asymptotic behavior of R<inf>n</inf>(x) defined by R<inf>n</inf>(x) = 1 - F*<sup>n</sup>(x) - n(1 - F(x)). Our results complement those previously obtained by several authors. In this paper, we define several new classes of functions related to regular variation and Oregular variation. As a typical result, in one of our theorems we show that R<inf>n</inf>(x) = O(l)f(x)R(x), where f(x) is the density of F and R(x) = ∫<inf>0</inf><sup>x</sup>(1 - F(y))dy. We also discuss some applications. © 1998 Plenum Publishing Corporation.
Año de publicación:
1998
Keywords:
Fuente:
scopusTipo de documento:
Article
Estado:
Acceso restringido
Áreas de conocimiento:
- Probabilidad
- Probabilidad
- Probabilidad
Áreas temáticas de Dewey:
- Probabilidades y matemática aplicada
- Matemáticas
- Álgebra
Objetivos de Desarrollo Sostenible:
- ODS 10: Reducción de las desigualdades
- ODS 16: Paz, justicia e instituciones sólidas
- ODS 17: Alianzas para lograr los objetivos