Estimation of Convolution Tails
Abstract
Several classes of distribution functions (d.f.) are originated by considering distributions whose tailfunctions satisfy special asymptotic relations. A large class sharing this property is provided by a certain subexponential class, in which case the asymptotic relation involves tails of convolution powers. This paper introduces a statistic which estimates the asymptotic behaviour of convolution tails of a given d.f. and it is shown that this statistic is strongly consistent and asymptotically normal under appropriate conditions. Furthermore, the statistic can be used to test the hypothesis that a d.f. is in the exponential class being described.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1987
- Accession Number
- ADA190323
Entities
People
- Eric Willekens
Organizations
- University of North Carolina at Chapel Hill