The maximum likelihood degree of mixtures of independence models

DOI10.1137/16M1088843MaRDI QIDQ5349294zbMATH OpenFDO

Authors Jose Israel Rodriguez, Botong Wang

Publication date 24 August 2017

Published in SIAM Journal on Applied Algebra and Geometry (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1505.06536

zbMATH Keywords

algebraic statisticsEuler characteristicsmaximum likelihood estimationmixture models

Mathematics Subject Classification ID

Applications of commutative algebra (e.g., to statistics, control theory, optimization, etc.) (13P25) Determinantal varieties (14M12) Point estimation (62F10) Applications of mathematical programming (90C90)

Abstract: The maximum likelihood degree (ML degree) measures the algebraic complexity of a fundamental optimization problem in statistics: maximum likelihood estimation. In this problem, one maximizes the likelihood function over a statistical model. The ML degree of a model is an upper bound to the number of local extrema of the likelihood function and can be expressed as a weighted sum of Euler characteristics. The independence model (i.e. rank one matrices over the probability simplex) is well known to have an ML degree of one, meaning their is a unique local maxima of the likelihood function. However, for mixtures of independence models (i.e. rank two matrices over the probability simplex), it was an open question as to how the ML degree behaved. In this paper, we use Euler characteristics to prove an outstanding conjecture by Hauenstein, the first author, and Sturmfels; we give recursions and closed form expressions for the ML degree of mixtures of independence models.

Recommendations

• The maximum likelihood degree
• Bounding the maximum likelihood degree
• Maximum likelihood degree of the two-dimensional linear Gaussian covariance model
• Maximum likelihood for matrices with rank constraints
• Maximum likelihood geometry in the presence of data zeros

Cites work

• \(D\)-modules, perverse sheaves, and representation theory. Translated from the Japanese by Kiyoshi Takeuchi
• A Vietoris Mapping Theorem for Homotopy
• Bounding the maximum likelihood degree
• Characteristic cycles for the loop Grassmannian and nilpotent orbits.
• Characteristic varieties and vanishing cycles
• Fixed points of the EM algorithm and nonnegative rank boundaries
• Geometry of maximum likelihood estimation in Gaussian graphical models
• scientific article; zbMATH DE number 3834955 (Why is no real title available?)
• scientific article; zbMATH DE number 52944 (Why is no real title available?)
• scientific article; zbMATH DE number 3235051 (Why is no real title available?)
• Introduction to Toric Varieties. (AM-131)
• Likelihood geometry
• Maximum likelihood degree of variance component models
• Maximum likelihood duality for determinantal varieties
• Maximum likelihood for dual varieties
• Maximum likelihood for matrices with rank constraints
• Maximum likelihood geometry in the presence of data zeros
• Sheaves in topology
• Solving the likelihood equations
• Statistics
• Tangents to an analytic variety
• The Gauss map and a noncompact Riemann-Roch formula for constructible sheaves on semiabelian varieties
• The maximum likelihood degree
• The maximum likelihood degree of a very affine variety
• The maximum likelihood degree of toric varieties
• The Signed Euler Characteristic of Very Affine Varieties

Cited in

(21)

• Defect of Euclidean distance degree
• Maximum likelihood for matrices with rank constraints
• Marginal likelihood integrals for mixtures of independence models
• Algebraic boundary of matrices of nonnegative rank at most three
• Decomposition of maximum likelihood in mixed graphical interaction models
• Maximum likelihood estimation for matrix normal models via quiver representations
• Reciprocal maximum likelihood degrees of diagonal linear concentration models
• The maximum likelihood degree of linear spaces of symmetric matrices
• A numerical approach for computing Euler characteristics of affine varieties
• Bounding the maximum likelihood degree
• The maximum likelihood degree
• Maximum likelihood degree, complete quadrics, and \(\mathbb{C}^*\)-action
• The Maximum Likelihood Degree of Sparse Polynomial Systems
• Algebraic structures in statistical methodology. Abstracts from the workshop held December 4--10, 2022
• Logarithmic cotangent bundles, Chern‐Mather classes, and the Huh‐Sturmfels involution conjecture
• Linear optimization on varieties and Chern-Mather classes
• Maximum likelihood estimation for unrooted 3-leaf trees: an analytic solution for the CFN model
• A new generic vanishing theorem on homogeneous varieties and the positivity conjecture for triple intersections of Schubert cells
• Applications of singularity theory in applied algebraic geometry and algebraic statistics
• Mixtures of discrete decomposable graphical models
• Boundary geometry, degeneracy, and identifiability transitions in binary latent star models

This page was built for publication: The maximum likelihood degree of mixtures of independence models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5349294)