Statements (39)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:algorithm
|
| gptkbp:application |
natural language processing
speech recognition bioinformatics image segmentation data imputation |
| gptkbp:category |
gptkb:logic
unsupervised learning |
| gptkbp:citation |
Maximum Likelihood from Incomplete Data via the EM Algorithm (1977)
|
| gptkbp:convergesTo |
local maximum of likelihood function
|
| gptkbp:field |
gptkb:machine_learning
statistics |
| gptkbp:form |
gptkb:Jensen's_inequality
likelihood function |
| gptkbp:fullName |
gptkb:Expectation-Maximization_algorithm
|
| https://www.w3.org/2000/01/rdf-schema#label |
EM algorithm
|
| gptkbp:input |
incomplete data
|
| gptkbp:introduced |
gptkb:Donald_Rubin
gptkb:Nan_Laird gptkb:Arthur_Dempster |
| gptkbp:introducedIn |
1977
|
| gptkbp:limitation |
slow convergence
sensitive to initial values may converge to local maxima |
| gptkbp:output |
parameter estimates
|
| gptkbp:purpose |
find maximum likelihood estimates of parameters in probabilistic models
|
| gptkbp:relatedTo |
gptkb:Baum-Welch_algorithm
gptkb:Gaussian_Mixture_Model gptkb:K-means_algorithm |
| gptkbp:step |
Expectation step
Maximization step |
| gptkbp:supportsAlgorithm |
iterative refinement
|
| gptkbp:usedFor |
clustering
parameter estimation mixture models latent variable models missing data problems |
| gptkbp:bfsParent |
gptkb:Computational_statistics
|
| gptkbp:bfsLayer |
6
|