Statements (52)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
statistical analysis
|
| gptkbp:assumes |
data points are generated from a mixture of Gaussians
|
| gptkbp:basedOn |
gptkb:Gaussian_distribution
|
| gptkbp:canBe |
data compression
image segmentation speech modeling anomaly detection background subtraction |
| gptkbp:canBeFittedBy |
gptkb:Markov_chain_Monte_Carlo
Bayesian inference maximum likelihood estimation variational inference |
| gptkbp:canBeRegularizedBy |
penalized likelihood
prior distributions |
| gptkbp:componentDistribution |
gptkb:univariate_normal_distribution
gptkb:multivariate_normal_distribution |
| gptkbp:estimatedCost |
expectation-maximization algorithm
|
| gptkbp:extendsTo |
Bayesian Gaussian mixture model
infinite Gaussian mixture model |
| gptkbp:field |
gptkb:signal_processing
computer vision speech recognition bioinformatics pattern recognition unsupervised learning |
| gptkbp:hasModel |
multimodal distributions
|
| https://www.w3.org/2000/01/rdf-schema#label |
Gaussian mixture model
|
| gptkbp:introduced |
gptkb:Karl_Pearson
|
| gptkbp:introducedIn |
1894
|
| gptkbp:limitation |
assumes Gaussian components
can get stuck in local optima sensitive to initialization |
| gptkbp:output |
cluster assignments
likelihood of data probability of cluster membership |
| gptkbp:parameter |
mean
covariance mixture weights |
| gptkbp:relatedTo |
Markov chain
k-means clustering finite mixture model soft clustering |
| gptkbp:usedFor |
density estimation
clustering |
| gptkbp:usedIn |
gptkb:machine_learning
statistics |
| gptkbp:visualizes |
scatter plots
contour plots |
| gptkbp:bfsParent |
gptkb:Gaussian_model
gptkb:Finite_mixture_models gptkb:Mixture_of_factor_analyzers |
| gptkbp:bfsLayer |
7
|