Statements (27)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:algorithm
statistical analysis |
| gptkbp:appliesTo |
gptkb:hidden_Markov_models
mixture models latent variable models |
| gptkbp:basedOn |
gptkb:Expectation-Maximization_algorithm
|
| gptkbp:category |
gptkb:logic
stochastic algorithm |
| gptkbp:convergesTo |
local maximum of likelihood
|
| gptkbp:handles |
intractable E-step
|
| https://www.w3.org/2000/01/rdf-schema#label |
Monte Carlo EM
|
| gptkbp:improves |
tractability of EM for complex models
|
| gptkbp:limitation |
computationally intensive
variance due to sampling |
| gptkbp:proposedBy |
1990
Wei and Tanner |
| gptkbp:relatedTo |
Stochastic EM
Variational EM |
| gptkbp:requires |
random sampling
|
| gptkbp:step |
E-step approximated by sampling
M-step maximizes expected log-likelihood |
| gptkbp:usedFor |
parameter estimation
|
| gptkbp:usedIn |
gptkb:machine_learning
statistics |
| gptkbp:uses |
gptkb:Monte_Carlo_integration
|
| gptkbp:bfsParent |
gptkb:expectation-maximization_algorithm
|
| gptkbp:bfsLayer |
7
|