Gaussian mixture models

GPTKB entity

Statements (52)
Predicate Object
gptkbp:instanceOf statistical analysis
Machine learning algorithm
gptkbp:alternativeTo gptkb:Dirichlet_process_mixture_models
gptkb:Finite_mixture_models
gptkbp:application gptkb:Bioinformatics
Speech recognition
Image segmentation
Anomaly detection
gptkbp:assumes Independence between components
Data is generated from a mixture of several Gaussian distributions
gptkbp:basedOn gptkb:Gaussian_distribution
gptkbp:canBe Diagonal covariance
Full covariance
Multivariate
Spherical covariance
Univariate
gptkbp:canBeRegularizedWith Covariance regularization
gptkbp:canBeTrainedWith Bayesian inference
Maximum likelihood estimation
gptkbp:category Generative model
Probabilistic model
Unsupervised learning
gptkbp:component Covariance matrix
Mean vector
Mixing coefficient
gptkbp:extendsTo gptkb:Variational_inference
gptkb:Mixture_of_experts
gptkb:Mixture_of_factor_analyzers
https://www.w3.org/2000/01/rdf-schema#label Gaussian mixture models
gptkbp:implementedIn gptkb:MATLAB
gptkb:TensorFlow_Probability
gptkb:scikit-learn
R
gptkbp:introduced gptkb:Karl_Pearson
gptkbp:introducedIn 1894
gptkbp:limitation Assumes Gaussianity
Can overfit with too many components
Sensitive to initialization
gptkbp:modelSelectionMethod gptkb:Akaike_Information_Criterion
gptkb:Bayesian_Information_Criterion
gptkbp:output Cluster assignments
Posterior probabilities
Probability densities
gptkbp:parameterEstimationMethod gptkb:Expectation-Maximization_algorithm
gptkbp:relatedTo gptkb:K-means_clustering
Hidden Markov models
Latent variable models
gptkbp:usedIn Clustering
Density estimation
gptkbp:visualizes Ellipses in 2D
gptkbp:bfsParent gptkb:Scikit-learn
gptkbp:bfsLayer 5