Computational Statistics

, Volume 29, Issue 3–4, pp 661–683 | Cite as

Holonomic gradient descent for the Fisher–Bingham distribution on the \(d\)-dimensional sphere

  • Tamio Koyama
  • Hiromasa Nakayama
  • Kenta Nishiyama
  • Nobuki Takayama
Original Paper

Abstract

We propose an accelerated version of the holonomic gradient descent and apply it to calculating the maximum likelihood estimate (MLE) of the Fisher–Bingham distribution on a \(d\)-dimensional sphere. We derive a Pfaffian system (an integrable connection) and a series expansion associated with the normalizing constant with an error estimation. These enable us to solve some MLE problems up to dimension \(d=7\) with a specified accuracy.

Keywords

Fisher–Bingham distribution Maximum likelihood estimate Holonomic gradient descent Integrable connection Pfaffian system 

References

  1. Bayer D, Stillman M (1988) On the complexity of computing syzygies. J Symb Comput 6:135–147CrossRefMATHMathSciNetGoogle Scholar
  2. Cano B, Stuart AM, Suli E, Warren JO (2001) Stiff oscillatory systems, delta jumps and white noise. Found Comput Math 1:69–99CrossRefMATHMathSciNetGoogle Scholar
  3. GSL.GNU scientific library. http://www.gnu.org/software/gsl/, 2012
  4. Hashiguchi H, Numata Y, Takayama N, Takemura A (2013) Holonomic gradient method for the distribution function of the largest root of Wishart matrix. J Multivar Anal 117:296–312CrossRefMATHMathSciNetGoogle Scholar
  5. Kent JT (1982) The Fisher-Bingham distribution on the sphere. J Roy Stat Soc Ser B 44:71–80MATHMathSciNetGoogle Scholar
  6. Koyama T (2013) A holonomic ideal which annihilates the Fisher–Bingham integral. Funkcialaj Ekvacioj 56:51–61CrossRefMATHMathSciNetGoogle Scholar
  7. Koyama T, Nakayama H, Nishiyama K, Takayama N (2012) The holonomic rank of the Fisher–Bingham system of differential equations. http://arxiv.org/abs/1205.6144
  8. Kume K, Walker SG (2009) On the Fisher–Bingham distribution. Stat Comput 19:167–172CrossRefMathSciNetGoogle Scholar
  9. Kume K, Wood ATA (2005) Saddlepoint approximations for the Bingham and Fisher–Bingham normalising constants. Biometrika 92:465–476CrossRefMATHMathSciNetGoogle Scholar
  10. Mardia KV, Jupp PE (2000) Directional statistics. Wiley, New YorkMATHGoogle Scholar
  11. Nakayama H, Nishiyama K, Noro M, Ohara K, Sei T, Takayama N, Takemura A (2011) Holonomic gradient descent and its application to the Fisher–Bingham integral. Adv Appl Math 47:639–658CrossRefMATHMathSciNetGoogle Scholar
  12. Nocedal J, Wright S (2007) Numerical optimization. Springer, New YorkGoogle Scholar
  13. Risa/Asir. A computer algebra system. http://www.math.kobe-u.ac.jp/Asir/, 2012
  14. Sei T, Shibata H, Takemura A, Ohara K, Takayama N (2013) Properties and applications of fisher distribution on the rotation group. J Multivar Anal 116:440–455CrossRefMATHMathSciNetGoogle Scholar
  15. Wood ATA (1988) Some notes on the Fisher–Bingham family on the sphere. Commun Stat Theory Methods 17:3881–3897CrossRefMATHGoogle Scholar
  16. Zeilberger D (1990) A holonomic systems approach to speical function identities. J Comput Appl Math 32:321–368CrossRefMATHMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Tamio Koyama
    • 1
  • Hiromasa Nakayama
    • 1
  • Kenta Nishiyama
    • 2
  • Nobuki Takayama
    • 1
  1. 1.Department of MathematicsKobe University and JST CREST Hibi TeamKobeJapan
  2. 2.School of Management and InformationUniversity of Shizuoka and JST CREST Hibi TeamShizuokaJapan

Personalised recommendations