Series approximation methods in statistics (Q1324805)

This book is focused on Edgeworth and saddlepoint approximations and related techniques not discussed in recent theoretical text books. Attention is paid to a rigorous consideration of Edgeworth and saddlepoint expansion limit theorems, especially for lattice distributions, and a survey of the more recent developments in this field. Saddlepoint methods of approximating densities and distribution functions for conditional distributions are investigated. As applications, approximation of densities and distribution functions of maximum likelihood estimators and likelihood ratio statistics are also considered. In the final chapter, computational tools are presented to help in the derivation and manipulation of these expansions.