Convergence results for the EM approach to mixtures of experts architectures

Neural Networks

Volumn 8, Issue 9, 1995, Pages 1409-1431

  Jordan, Michael I ^a   Xu, Lei ^a  

^a MASSACHUSETTS INSTITUTE OF TECHNOLOGY   (United States)

Author keywords

Convergence rate; EM algorithm; Hierarchical models; Maximum likelihood; Mixture models; Optimization; Statistical models; Supervised learning

Indexed keywords

COMPUTER ARCHITECTURE; COMPUTER SIMULATION; CONVERGENCE OF NUMERICAL METHODS; EXPERT SYSTEMS; ITERATIVE METHODS; LEARNING SYSTEMS; NEURAL NETWORKS; NUMERICAL ANALYSIS; OPTIMIZATION; PARAMETER ESTIMATION;

CONVERGENCE RATE; EXPECTATION MAXIMIZATION ALGORITHM; EXPERT ARCHITECTURE; HIERARCHICAL MIXTURE; STATISTICAL MODEL;

ALGORITHMS;

ALGORITHM; ARTICLE; ARTIFICIAL NEURAL NETWORK; LEARNING; PRIORITY JOURNAL;

EID: 0029617280     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/0893-6080(95)00014-3     Document Type: Article

References (21)

1. Growth transformation for functions on manifolds
2. A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains
3. Parameter estimation for a mixture of linear regressions
4. Maximum-likelihood from incomplete data via the EM algorithm
5. Multivariate adaptive regression splines
6. Adaptive mixtures of local experts
7. Hierarchies of adaptive experts
8. Hierarchical mixtures of experts and the EM algorithm
9. A fast improvement to the EM algorithm on its own terms
10. An iterative procedure for obtaining maximum-likelihood estimates of the parameters for a mixture of normal distributions
11. The numerical evaluation of the maximum-likelihood estimates of a subset of mixture proportions
12. A new approach to estimating switching regressions
13. Estimating mixtures of normal distribution and switching regressions
14. Mixture densities, maximum likelihood, and the EM algorithm
15. On the convergence properties of the EM algorithm
16. Theoretical and experimental studies of the EM algorithm for unsupervised learning based on finite Gaussian mixtures
17. On convergence properties of the EM algorithm for Gaussian mixtures