Laplace approximations to hypergeometric functions of two matrix arguments

JOURNAL OF MULTIVARIATE ANALYSIS

卷:94

Article

Butler, RW ; Wood, ATA

关键词: asymptotic approximation;    eigenvalue distribution;    matrix-argument hypergeometric function;    laplace approximation;   

DOI  :  10.1016/j.jmva.2004.05.010

来源: Elsevier

【 摘 要 】

We present a unified approach to Laplace approximation of hypergeometric functions with two matrix arguments. The general form of the approximation is designed to exploit the Laplace approximations to hypergeometric functions of a single matrix argument presented in Butler and Wood (Ann. Statist. 30 (2002) 1155, Laplace approximations to Bessel functions of matrix argument, J. Comput. Appl. Math. 155 (2003) 359) which have proved to be very accurate in a variety of settings. All but one of the approximations presented here appear to be new. Numerical accuracy is investigated in a number of statistical applications. (c) 2004 Elsevier Inc. All rights reserved.

【 授权许可】

Free