You are here

Home » Content

Moment Varieties of Gaussian Mixtures

Journal Name:

  • Journal of Algebraic Statistics

Publication Year:

  • 2016

Key Words:

Author NameUniversity of Author
Abstract (2. Language): 
The points of a moment variety are the vectors of all moments up to some order, for a given family of probability distributions. We study the moment varieties for mixtures of multivariate Gaussians. Following up on Pearson's classical work from 1894, we apply current tools from computational algebra to recover the parameters from the moments. Our moment varieties extend objects familiar to algebraic geometers. For instance, the secant varieties of Veronese varieties are the loci obtained by setting all covariance matrices to zero. We compute the ideals of the 5-dimensional moment varieties representing mixtures of two univariate Gaussians, and we o er a comparison to the maximum likelihood approach.
Bookmark/Search this post with
FULL TEXT (PDF): 
PDF icon arastirmax-moment-varieties-gaussian-mixtures.pdf
  • 1
14
28
  • English

REFERENCES

References: 

[1] Carlos Amendola, Mathias Drton and Bernd Sturmfels: Maximum likelihood estimates for
Gaussian mixtures are transcendental, Mathematical Aspects of Computer and Information
Sciences, (eds. Ilias Kotsireas, Siegfried Rump and Chee Yap), MACIS 2015, Berlin, (2016)
pp. 579{590.
[2] Anima Anandkumar, Rong Ge, Daniel Hsu, Sham Kakade and Matus Telgarsky: Tensor
decompositions for learning latent variable models, J. Mach. Learn. Res. 15 (2014) 2773{
2832.
REFERENCES 28
[3] Mikhail Belkin and Kaushik Sinha: Polynomial learning of distribution families, FOCS
2010, 51st Annual IEEE Symposium on Foundations of Computer Science, 2010, pp. 103{
112.
[4] Maria Chiara Brambilla and Giorgio Ottaviani: On the Alexander-Hirschowitz theorem, J.
Pure Appl. Algebra 212 (2008) 1229{1251.
[5] Ciro Ciliberto, Maria Angelica Cueto, Massimiliano Mella, Kristian Ranestad and Piotr
Zwiernik: Cremona linearizations of some classical varieties, arXiv:1403.1814, `Homage
to Corrado Segre" (Torino, Italy, 2014), Trends in the History of Science, Birkhauser, Basel.
[6] Mathias Drton, Bernd Sturmfels and Seth Sullivant: Lectures on Algebraic Statistics, Oberwolfach
Seminars, Vol 39, Birkhauser, Basel, (2009).
[7] Francesco Faa Di Bruno: Theorie des Formes Binaires, Brero Succr. de P. Marietti, Turin,
1876.
[8] David Eisenbud: Commutative Algebra: with a view toward Algebraic Geometry, Graduate
Texts in Mathematics 150, Springer, New York (1995).
[9] Rong Ge, Qingqing Huang and Sham Kakade: Learning mixtures of Gaussians in high
dimensions, STOC 2015, Proceedings of the 47th ACM on Symposium on Theory of Computing,
New York, pp. 761-770.
[10] Daniel Lazard: Injectivity of real rational mappings: the case of a mixture of two Gaussian
laws, Mathematics and Computers in Simulation 67 (2004) 67{84.
[11] Peter McCullagh: Tensor Methods in Statistics, Monographs on Statistics and Applied
Probability, Chapman & Hall, London, 1987.
[12] Karl Pearson: Contributions to the mathematical theory of evolution, Phil. Trans. Roy. Soc.
London A 185 (1894) 71{110.
[13] Sidney Yakowitz and John Spragins: On the identiability of nite mixtures, Ann. Math.
Statist. 39 (1968) 209{214.

Thank you for copying data from http://www.arastirmax.com