Abstract
In this paper we implement the holonomic gradient method to exactly compute the normalising constant of Bingham distributions. This idea is originally applied for general Fisher–Bingham distributions in Nakayama et al. (Adv. Appl. Math. 47:639–658, 2011). In this paper we explicitly apply this algorithm to show the exact calculation of the normalising constant; derive explicitly the Pfaffian system for this parametric case; implement the general approach for the maximum likelihood solution search and finally adjust the method for degenerate cases, namely when the parameter values have multiplicities.
Similar content being viewed by others
References
Barndorff-Nielsen, O.: Information and Exponential Families. Wiley Series in Probability and Mathematical Statistics. Wiley, New York (1978)
Feller, W.: An Introduction to Probability Theory and Its Applications vol. 2. Wiley, New York (1971)
Hashiguchi, H., Numata, Y., Takayama, N., Takemura, A.: Holonomic gradient method for the distribution function of the largest root of a Wishart matrix. J. Multivar. Anal. 117, 296–312 (2013)
Kent, J.T.: The complex Bingham distribution and shape analysis. J. R. Stat. Soc. 56, 285–289 (1994)
Koyama, T.: A holonomic ideal annihilating the Fisher–Bingham integral (2011). arXiv:1104.1411
Koyama, T., Nakayama, H., Nishiyama, K., Takayama, N.: Holonomic gradient descent for the Fisher–Bingham distribution on the n-dimensional sphere (2012a). arXiv:1201.3239
Koyama, T., Nakayama, H., Nishiyama, K., Takayama, N.: The holonomic rank of the Fisher–Bingham system of differential equations (2012b). arXiv:1205.6144
Kume, A., Wood, A.T.A.: Saddlepoint approximations for the Bingham and Fisher–Bingham normalising constants. Biometrika 92, 465–476 (2005)
Kume, A., Wood, A.T.A.: On the derivatives of the normalising constant of the Bingham distribution. Stat. Probab. Lett. 77, 832–837 (2007)
Mardia, K.V., Jupp, P.E.: Directional Statistics. Wiley Series in Probability and Statistics. Wiley, Chichester (2000). ISBN 0-471-95333-4
Nakayama, H., Nishiyama, K., Noro, M., Ohara, K., Sei, T., Takayama, N., Takemura, A.: Holonomic gradient descent and its application to the Fisher–Bingham integral. Adv. Appl. Math. 47, 639–658 (2011)
Sei, T., Shibata, H., Takemura, A., Ohara, K., Takayama, N.: Properties and applications of Fisher distribution on the rotation group. J. Multivar. Anal. 116, 440–455 (2013)
Sun, W., Yuan, Y.-X.: Optimization Theory and Methods. Springer, New York (2006)
Takayama, N.: Propagation of singularities of solutions of the Euler–Darboux equation and a global structure of the space of holonomic functions I. Funkc. Ekvacioj 35, 343–403 (1992)
Acknowledgements
The first author is supported by JSPS Institutional Program for Young Researcher Overseas Visits.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A: One-dimensional representation
First we briefly describe derivation of the one-dimensional representation (3) according to Kume and Wood (2005). Note that the parameter λ i in their paper is our −θ i . Consider p independent normal random variables x i ∼N(0,(−2θ i )−1), where θ i <0 for all i. Then the marginal density of \(r=\sum_{i=1}^{p}x_{i}^{2}\) is directly calculated as
On the other hand, the characteristic function of r is
In general, the density function is represented by its characteristic function as
at arbitrary continuous point of f (see e.g. Feller 1971). By combining Eqs. (26) to (28), we have
Note that this expression holds for any p≥2. If p≥3, then
since \(|\prod_{k=1}^{p}(-\theta_{k}-is)^{-1/2}|\) is integrable over (−∞,∞). By analytic continuation with respect to s, we obtain
for any real number t 0 less than min k (−θ k ). Even if some θ i ’s are not negative, Eq. (29) still holds due to analytic continuation with respect to θ, as long as t 0<min k (−θ k ). Hence we obtain (3).
Appendix B: Truncation error of the power series
We derive the power series expansion of c(θ) and evaluate the truncation error according to Koyama et al. (2012a). Let θ(ϕ,d)=(ϕ 1,…,ϕ 1,…,ϕ q ,…,ϕ q ) be a parameter vector with multiplicities d=(d 1,…,d q ). By Kume and Wood (2007), we have
Let
Then the truncation error is evaluated as
Rights and permissions
About this article
Cite this article
Sei, T., Kume, A. Calculating the normalising constant of the Bingham distribution on the sphere using the holonomic gradient method. Stat Comput 25, 321–332 (2015). https://doi.org/10.1007/s11222-013-9434-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-013-9434-0