Improved initialisation of model-based clustering using Gaussian hierarchical partitions
From MaRDI portal
Abstract: Initialisation of the EM algorithm in model-based clustering is often crucial. Various starting points in the parameter space often lead to different local maxima of the likelihood function and, so to different clustering partitions. Among the several approaches available in the literature, model-based agglomerative hierarchical clustering is used to provide initial partitions in the popular MCLUST R package. This choice is computationally convenient and often yields good clustering partitions. However, in certain circumstances, poor initial partitions may cause the EM algorithm to converge to a local maximum of the likelihood function. We propose several simple and fast refinements based on data transformations and illustrate them through data examples.
Recommendations
- Improved model-based clustering performance using Bayesian initialization averaging
- Initializing the EM algorithm in Gaussian mixture models with an unknown number of components
- A Gaussian mixture model based \(k\)-means to initialize the EM algorithm
- Initializing the EM algorithm for univariate Gaussian, multi-component, heteroscedastic mixture models by dynamic programming partitions
- A robust EM clustering algorithm for Gaussian mixture models
Cites work
- scientific article; zbMATH DE number 41467 (Why is no real title available?)
- scientific article; zbMATH DE number 3567782 (Why is no real title available?)
- scientific article; zbMATH DE number 1348600 (Why is no real title available?)
- scientific article; zbMATH DE number 1070609 (Why is no real title available?)
- Algorithms for Model-Based Gaussian Hierarchical Clustering
- Choosing starting values for the EM algorithm for getting the highest likelihood in multivariate Gaussian mixture models
- Cluster Analysis
- Finding Groups in Data
- Finite mixture models
- Finite mixture models and model-based clustering
- How Many Clusters? Which Clustering Method? Answers Via Model-Based Cluster Analysis
- Initializing the EM algorithm in Gaussian mixture models with an unknown number of components
- Model-Based Clustering, Classification, and Density Estimation Using mclust in R
- Model-Based Clustering, Discriminant Analysis, and Density Estimation
- Model-Based Gaussian and Non-Gaussian Clustering
- Model-based cluster and discriminant analysis with the MIXMOD software
- On the convergence properties of the EM algorithm
- The EM Algorithm and Extensions, 2E
- Variable Selection for Model-Based Clustering
Cited in
(14)- Improved model-based clustering performance using Bayesian initialization averaging
- Addressing overfitting and underfitting in Gaussian model-based clustering
- Practical Initialization of Recursive Mixture-Based Clustering for Non-negative Data
- Better than the best? Answers via model ensemble in density-based clustering
- A stochastic block model for interaction lengths
- An artificial bee colony algorithm for mixture model-based clustering
- Constrained clustering with a complex cluster structure
- Model-based clustering with sparse covariance matrices
- Algorithms for Model-Based Gaussian Hierarchical Clustering
- Modelling the role of variables in model-based cluster analysis
- Gaussian model-based partitioning using iterated local search
- Group-wise shrinkage estimation in penalized model-based clustering
- Unobserved classes and extra variables in high-dimensional discriminant analysis
- Estimation and Testing Problems in Auditory Neuroscience via Clustering
This page was built for publication: Improved initialisation of model-based clustering using Gaussian hierarchical partitions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2418409)
