Data-Discriminants of Likelihood Equations

Maximum likelihood estimation (MLE) is a fundamental computational problem in statistics. The problem is to maximize the likelihood function with respect to given data on a statistical model. An algebraic approach to this problem is to solve a very structured parameterized polynomial system called likelihood equations. For general choices of data, the number of complex solutions to the likelihood equations is finite and called the ML-degree of the model. The only solutions to the likelihood equations that are statistically meaningful are the real/positive solutions. However, the number of real/positive solutions is not characterized by the ML-degree. We use discriminants to classify data according to the number of real/positive solutions of the likelihood equations. We call these discriminants data-discriminants (DD). We develop a probabilistic algorithm for computing DDs. Experimental results show that, for the benchmarks we have tried, the probabilistic algorithm is more efficient than the standard elimination algorithm. Based on the computational results, we discuss the real root classification problem for the 3 by 3 symmetric matrix~model.

[1]  Elizabeth Gross,et al.  Maximum likelihood degree of variance component models , 2011, 1111.3308.

[2]  T. Willmore.  Algebraic Geometry , 1973, Nature.

[3]  Fabrice Rouillier,et al.  Solving parametric polynomial systems , 2004, J. Symb. Comput..

[4]  Jose Israel Rodriguez,et al.  Maximum likelihood for dual varieties , 2014, SNC.

[5]  Seth Sullivant,et al.  Lectures on Algebraic Statistics , 2008 .

[6]  Zbigniew Jelonek,et al.  Testing sets for properness of polynomial mappings , 1999 .

[7]  L. Pachter,et al.  Algebraic Statistics for Computational Biology: Preface , 2005 .

[8]  Éric Schost,et al.  Properness Defects of Projections and Computation of at Least One Point in Each Connected Component of a Real Algebraic Set , 2004, Discret. Comput. Geom..

[9]  D. Mumford.  Algebraic Geometry I: Complex Projective Varieties , 1981 .

[10]  Seth Sullivant,et al.  Algebraic and Geometric Methods in Statistics: The algebraic complexity of maximum likelihood estimation for bivariate missing data , 2009 .

[11]  George E. Collins,et al.  Partial Cylindrical Algebraic Decomposition for Quantifier Elimination , 1991, J. Symb. Comput..

[12]  Mohab Safey El Din,et al.  Variant quantifier elimination , 2012, J. Symb. Comput..

[13]  Éric Schost,et al.  Computing Parametric Geometric Resolutions , 2003, Applicable Algebra in Engineering, Communication and Computing.

[14]  Bican Xia,et al.  A complete algorithm for automated discovering of a class of inequality-type theorems , 2001, Science in China Series F Information Sciences.

[15]  Bernd Sturmfels,et al.  The maximum likelihood degree , 2004, math/0406533.

[16]  David A. Cox,et al.  Ideals, Varieties, and Algorithms , 1997 .

[17]  Jonathan D. Hauenstein.  Maximum Likelihood for Matrices with Rank Constraints , 2014 .

[18]  Mohab Safey El Din,et al.  Probabilistic Algorithm for Polynomial Optimization over a Real Algebraic Set , 2013, SIAM J. Optim..

[19]  Masahiro Shiota,et al.  Nash triviality in families of Nash manifolds , 1992 .

[20]  Andrew J. Sommese,et al.  The numerical solution of systems of polynomials - arising in engineering and science , 2005 .

[21]  Changbo Chen,et al.  Triangular decomposition of semi-algebraic systems , 2013, J. Symb. Comput..

[22]  Bernd Sturmfels,et al.  Likelihood Geometry , 2013, 1305.7462.

[23]  Éric Schost,et al.  Polar varieties and computation of one point in each connected component of a smooth real algebraic set , 2003, ISSAC '03.

[24]  W. C. Mcginnis.  Ideals , 1925, Free Speech.

[25]  Donal O'Shea,et al.  Ideals, varieties, and algorithms - an introduction to computational algebraic geometry and commutative algebra (2. ed.) , 1997, Undergraduate texts in mathematics.

[26]  Jose Israel Rodriguez,et al.  Maximum likelihood geometry in the presence of data zeros , 2013, ISSAC.

[27]  Jean-Charles Faugère,et al.  Critical points and Gröbner bases: the unmixed case , 2012, ISSAC.

[28]  Marc Giusti,et al.  A Gröbner Free Alternative for Polynomial System Solving , 2001, J. Complex..

[29]  Donald St. P. Richards,et al.  Counting and locating the solutions of polynomial systems of maximum likelihood equations, I , 2006, J. Symb. Comput..

[30]  S. Basu,et al.  Algorithms in real algebraic geometry , 2003 .

[31]  Bernd Sturmfels,et al.  Solving the Likelihood Equations , 2005, Found. Comput. Math..

[32]  Caroline Uhler,et al.  Geometry of maximum likelihood estimation in Gaussian graphical models , 2010, 1012.2643.