Abstract
With the df F of the rv X we associate the natural exponential family of df's Fλ where dFλ (x)=eλx dF(x)/EeλX for λ∈Λ:={λ∈ℝ EeλX<∞}. Assume λ∞=sup Λ≤∞ does not lie in Λ. Let λ↑λ∞, then non-degenerate limit laws for the normalised distributions Fλ(aλx+bλ) are the normal and gamma distributions. Their domains of attractions are determined. Applications to saddlepoint and gamma approximations are considered.
Original language | English |
---|---|
Pages (from-to) | 83-103 |
Number of pages | 21 |
Journal | Stochastic Processes and their Applications |
Volume | 107 |
Issue number | 1 |
DOIs | |
State | Published - 1 Sep 2003 |
Keywords
- Asymptotic normality
- Convex conjugate
- Domain of attraction
- Esscher transform
- Exponential family
- Gamma distribution
- Laplace transform
- Normal distribution
- Regular variation
- Self-neglecting
- Strongly unimodal