Inference and Mixture Modeling with the Elliptical Gamma Distribution
arXiv:1410.4812 · doi:10.1016/j.csda.2016.02.009
Abstract
We study modeling and inference with the Elliptical Gamma Distribution (EGD). We consider maximum likelihood (ML) estimation for EGD scatter matrices, a task for which we develop new fixed-point algorithms. Our algorithms are efficient and converge to global optima despite nonconvexity. Moreover, they turn out to be much faster than both a well-known iterative algorithm of Kent & Tyler (1991) and sophisticated manifold optimization algorithms. Subsequently, we invoke our ML algorithms as subroutines for estimating parameters of a mixture of EGDs. We illustrate our methods by applying them to model natural image statistics---the proposed EGD mixture model yields the most parsimonious model among several competing approaches.
arXiv comment23 pages, 11 figures
References in corpus (9)
- Manopt, a Matlab toolbox for optimization on manifolds
- A geometric analysis of subspace clustering with outliers
- Conic geometric optimisation on the manifold of positive definite matrices
- Breakdown points for maximum likelihood estimators of location-scale mixtures
- Multivariate Generalized Gaussian Distribution: Convexity and Graphical Models
- Robust recovery of multiple subspaces by geometric l_p minimization
- In All Likelihood, Deep Belief Is Not Enough
- New Algorithms for -Estimation of Multivariate Scatter and Location
- MixEst: An Estimation Toolbox for Mixture Models