Model-based clustering with sparse covariance matrices
From MaRDI portal
Publication:2329799
Abstract: Finite Gaussian mixture models are widely used for model-based clustering of continuous data. Nevertheless, since the number of model parameters scales quadratically with the number of variables, these models can be easily over-parameterized. For this reason, parsimonious models have been developed via covariance matrix decompositions or assuming local independence. However, these remedies do not allow for direct estimation of sparse covariance matrices nor do they take into account that the structure of association among the variables can vary from one cluster to the other. To this end, we introduce mixtures of Gaussian covariance graph models for model-based clustering with sparse covariance matrices. A penalized likelihood approach is employed for estimation and a general penalty term on the graph configurations can be used to induce different levels of sparsity and incorporate prior knowledge. Model estimation is carried out using a structural-EM algorithm for parameters and graph structure estimation, where two alternative strategies based on a genetic algorithm and an efficient stepwise search are proposed for inference. With this approach, sparse component covariance matrices are directly obtained. The framework results in a parsimonious model-based clustering of the data via a flexible model for the within-group joint distribution of the variables. Extensive simulated data experiments and application to illustrative datasets show that the method attains good classification performance and model quality.
Recommendations
- Model-based clustering based on sparse finite Gaussian mixtures
- Using conditional independence for parsimonious model-based Gaussian clustering
- Penalized model-based clustering with unconstrained covariance matrices
- Sparse covariance estimation in heterogeneous samples
- Sparse Gaussian graphical mixture model
Cites work
- A genetic algorithm for graphical model selection
- A macro-DAG structure based mixture model
- Adaptive Model Selection
- Ancestral graph Markov models.
- Bayesian inference of multiple Gaussian graphical models
- Bayesian model averaging: A tutorial. (with comments and a rejoinder).
- Bayesian regularization for normal mixture estimation and model-based clustering
- Bayesian sparse graphical models for classification with application to protein expression data
- Convergence Criteria for Genetic Algorithms
- Covariance chains
- Covariance estimation: the GLM and regularization perspectives
- Covariance matrix selection and estimation via penalised normal likelihood
- EM for mixtures
- Estimating the dimension of a model
- Estimation of a covariance matrix with zeros
- Estimation of multiple networks in Gaussian mixture models
- Extended Bayesian information criteria for model selection with large model spaces
- Finite mixture and Markov switching models.
- Finite mixture models
- Genetic algorithms and their statistical applications: an introduction
- High-dimensional covariance estimation based on Gaussian graphical models
- High-dimensional Ising model selection with Bayesian information criteria
- scientific article; zbMATH DE number 3126094 (Why is no real title available?)
- scientific article; zbMATH DE number 3150484 (Why is no real title available?)
- scientific article; zbMATH DE number 3168330 (Why is no real title available?)
- scientific article; zbMATH DE number 4211299 (Why is no real title available?)
- scientific article; zbMATH DE number 4159863 (Why is no real title available?)
- scientific article; zbMATH DE number 3942813 (Why is no real title available?)
- scientific article; zbMATH DE number 3567782 (Why is no real title available?)
- scientific article; zbMATH DE number 194544 (Why is no real title available?)
- scientific article; zbMATH DE number 1906319 (Why is no real title available?)
- scientific article; zbMATH DE number 922429 (Why is no real title available?)
- Hyper Inverse Wishart Distribution for Non-decomposable Graphs and its Application to Bayesian Inference for Gaussian Graphical Models
- Improved initialisation of model-based clustering using Gaussian hierarchical partitions
- Introduction to Graphical Modelling
- Joint estimation of multiple graphical models
- Joint structural estimation of multiple graphical models
- Model Selection and Accounting for Model Uncertainty in Graphical Models Using Occam's Window
- Model selection and estimation in the Gaussian graphical model
- Model-based clustering
- Model-based clustering based on sparse finite Gaussian mixtures
- Model-based clustering of high-dimensional data: a review
- Model-Based Clustering, Discriminant Analysis, and Density Estimation
- Model-Based Gaussian and Non-Gaussian Clustering
- Modelling the role of variables in model-based cluster analysis
- Node-based learning of multiple Gaussian graphical models
- On the ``degrees of freedom of the lasso
- On the convergence properties of the EM algorithm
- Penalized Maximum Likelihood Estimator for Normal Mixtures
- Penalized model-based clustering with application to variable selection
- Penalized model-based clustering with unconstrained covariance matrices
- Positive definite estimators of large covariance matrices
- Probabilistic graphical models.
- Random graphs.
- Regularized parameter estimation in high-dimensional Gaussian mixture models
- Scaling it up: stochastic search structure learning in graphical models
- Simultaneous model-based clustering and visualization in the Fisher discriminative subspace
- Sparse covariance estimation in heterogeneous samples
- Sparse estimation of a covariance matrix
- Sparse Gaussian graphical mixture model
- Sparse inverse covariance estimation with the graphical lasso
- Stable and visualizable Gaussian parsimonious clustering models
- Statistical analysis of finite mixture distributions
- Structural pursuit over multiple undirected graphs
- Technological modelling for graphical models: an approach based on genetic algorithms
- The Joint Graphical Lasso for Inverse Covariance Estimation Across Multiple Classes
- Using conditional independence for parsimonious model-based Gaussian clustering
- Variable Selection for Clustering with Gaussian Mixture Models
- Variable Selection for Model-Based Clustering
- Variable Selection in Penalized Model‐Based Clustering Via Regularization on Grouped Parameters
- Variable selection methods for model-based clustering
Cited in
(23)- Sparse covariance estimation in heterogeneous samples
- On parsimonious models for modeling matrix data
- High-dimensional structure learning of sparse vector autoregressive models using fractional marginal pseudo-likelihood
- Gaussian mixture model with an extended ultrametric covariance structure
- Model-based clustering with determinant-and-shape constraint
- From here to infinity: sparse finite versus Dirichlet process mixtures in model-based clustering
- Group-wise shrinkage estimation in penalized model-based clustering
- Models of random sparse eigenmatrices and Bayesian analysis of multivariate structure
- Graphical models for sparse data: graphical Gaussian models with vertex and edge symmetries
- Reduced-Rank Modeling for High-Dimensional Model-Based Clustering
- Regularized matrix data clustering and its application to image analysis
- Clustered sparse structural equation modeling for heterogeneous data
- A bootstrap-augmented alternating expectation-conditional maximization algorithm for mixtures of factor analyzers
- Improving model choice in classification: an approach based on clustering of covariance matrices
- Conditional mixture modelling for heavy-tailed and skewed data
- Parsimonious ultrametric Gaussian mixture models
- The parsimonious Gaussian mixture models with partitioned parameters and their application in clustering
- Model-based clustering based on sparse finite Gaussian mixtures
- Parsimonious Gaussian mixture models with piecewise-constant eigenvalue profiles
- Sparse Model-Based Clustering of Three-Way Data via Lasso-Type Penalties
- Model-based clustering for covariance matrices via penalized Wishart mixture models
- Using conditional independence for parsimonious model-based Gaussian clustering
- Gaussian parsimonious clustering models with covariates and a noise component
This page was built for publication: Model-based clustering with sparse covariance matrices
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2329799)
