We extend recent work on Laplace approximations (Tierney and Kadane 1986; Tierney, Kass, and Kadane 1989) from parameter spaces that are subspaces of Rk to those that are on circles, spheres, and cylinders. While such distributions can be mapped onto the real line (for example, a distribution on the circle can be thought of as a function of an angle θ, 0 ⩽ 0 ⩾ 2π), that the end points coincide is not a feature of the real line, and requires special treatment. Laplace approximations on the real line make essential use of the normal integral in both the numerator and the denominator. Here that role is played by the von Mises integral on the circle, by the Bingham integrals on the spheres and hyperspheres, and by the normal-von Mises and normal-Bingham integrals on the cylinders and hypercylinders, respectively. We begin with a brief introduction to Laplace approximations and to previous Bayesian work on circles, spheres, and cylinders. We then develop the theory for parameter spaces that are hypercylinders, since all other shapes considered here are special cases. We compute some examples, which show reasonable accuracy even for small samples.
Des travaux recents sur les approximations de Laplace (Tierney et Kadane 1986; Tierney, Kass et Kadane 1989) dans le contexte d'espaces parametriques qui sont des sous-espaces de Rk, sont generalises a des espaces parametriques circulates, spheriques ou cylindriques. Měme si les lois considerees peuvent ětre transposees sur la droite reelle (par exemple, une loi definie sur le cercle peut ětre consideree comme une fonction d'un angle θ, 0 < 0 < 2π), le fait que les extremites coincident n'est pas une caracteristique de la droite reelle et demande un traitement particulier. Les approximations de Laplace sur la droite reelle utilisent l'integrate de la densite d'une loi normale (au numerateur et au denominateur). Ici ce rǒle est joue par les integrates de la densite de von Mises sur le cercle, de Bingham sur les spheres et les hyperspheres et les combinaisons normale-von Mises, normal-Bingham sur les cylindres et hypercylindres respectivement. Une courte revue sur les approximations de Laplace et les travaux anterieurs en statistique bayesienne dans le contexte des cercles, des spheres et des cylindres, est presentee. La theorie pour des espaces parametriques hypercylindriques est developpee; toutes les autres formes sont des cas particuliers de celle-ci. Des calculs effectues pour quelques exemples indiquent que la precision est raisonnable, měme pour de petits echantillons.
[1]
L. Tierney,et al.
Accurate Approximations for Posterior Moments and Marginal Densities
,
1986
.
[2]
J. E. H. Shaw,et al.
The implementation of the bayesian paradigm
,
1985
.
[3]
A. Lo,et al.
Bayes procedures for rotationally symmetric models on the sphere
,
1987
.
[4]
Calyampudi R. Rao,et al.
Kernel estimators of density function of directional data
,
1988
.
[5]
Peter E. Jupp,et al.
A Unified View of the Theory of Directional Statistics, 1975-1988
,
1989
.
[6]
S. Stigler.
Laplace's 1774 Memoir on Inverse Probability
,
1986
.
[7]
Adrian F. M. Smith,et al.
Sampling-Based Approaches to Calculating Marginal Densities
,
1990
.
[8]
R. Fisher.
Dispersion on a sphere
,
1953,
Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences.
[9]
K. J. Arnold.
On spherical probability distributions
,
1941
.
[10]
L. Tierney,et al.
Fully Exponential Laplace Approximations to Expectations and Variances of Nonpositive Functions
,
1989
.
[11]
Herman K. van Dijk,et al.
An algorithm for the computation of posterior moments and densities using simple importance sampling
,
1986
.
[12]
Irwin Guttman,et al.
Theoretical considerations of the multivariate von Mises-Fisher distribution
,
1988
.
[13]
A. Constantine.
Some Non-Central Distribution Problems in Multivariate Analysis
,
1963
.
[14]
Kanti V. Mardia,et al.
Bayesian inference for the von Mises-Fisher distribution
,
1976
.
[15]
Kanti V. Mardia,et al.
Statistics of Directional Data
,
1972
.
[16]
P. Guttorp,et al.
Finding the Location of a Signal: A Bayesian Analysis
,
1988
.
[17]
A. James.
Normal Multivariate Analysis and the Orthogonal Group
,
1954
.