A mixture of common skew‐t factor analysers
StatPublished 1 March 2014Open access
Paula M. Murray, Paul D. McNicholas, Ryan P. Browne
Citations42
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This paper is the first time that skewed common factors have been used, and it marks an important step in robust clustering and classification of high‐dimensional data.
Abstract
Abstract A mixture of common skew‐ t factor analysers model is introduced for model‐based clustering of high‐dimensional data. By assuming common factors, this model allows clustering to be performed in the presence of a large number of mixture components or when the number of dimensions is too large to be well modelled by the mixture of factor analysers model or a variant thereof. Furthermore, assuming that the component densities follow a skew‐ t distribution allows robust clustering of data with asymmetric clusters. This paper is the first time that skewed common factors have been used, and it marks an important step in robust clustering and classification of high‐dimensional data. The alternating expectation–conditional maximization algorithm is employed for parameter estimation. We demonstrate excellent clustering performance when our mixture of common skew‐ t factor analysers model is applied to real and simulated data. Copyright © 2014 John Wiley & Sons, Ltd
Keywords
Computer ScienceMathematics
Journal of the Royal Statistical Society Series B (Statistical Methodology)Maximum Likelihood from Incomplete Data Via the <i>EM</i> Algorithm
49,657 Citations1977A. P. Dempster, N. M. Laird +1 more
Journal of ClassificationComparing partitions
7,714 Citations1985Lawrence J. Hubert, Phipps Arabie
A measure based on the comparison of object triples having the advantage of a probabilistic interpretation in addition to being corrected for chance is proposed and bounded between ±1.5 and ±2.5.
Journal of the Royal Statistical Society Series B (Statistical Methodology)Non-Gaussian Ornstein–Uhlenbeck-based Models and Some of Their Uses in Financial Economics
1,890 Citations2001Ole E. Barndorff–Nielsen, Neil Shephard
BiometrikaMaximum likelihood estimation via the ECM algorithm: A general framework
1,814 Citations1993Xiao‐Li Meng, Donald B. Rubin
Nature GeneticsMLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia
1,711 Citations2001Scott A. Armstrong, Jane Staunton +8 more
It is proposed that MLL constitute a distinct disease, denoted here as MLL, and it is shown that the differences in gene expression are robust enough to classify leukemias correctly as M LL, acute lymphoblastic leukemia or acute myelogenous leukemia.
Journal of the Royal Statistical Society Series B (Statistical Methodology)The EM Algorithm—an Old Folk-song Sung to a Fast New Tune
807 Citations1997Xiao‐Li Meng, David A. van Dyk
A general alternating expectation–conditional maximization algorithm AECM is formulated that couples flexible data augmentation schemes with model reduction schemes to achieve efficient computations and shows the potential for a dramatic reduction in computational time with little increase in human effort.
Proceedings of the Royal Society of EdinburghXXV.—On Bernoulli's Numerical Solution of Algebraic Equations
542 Citations1927A. C. Aitken
BMC BioinformaticsClustering cancer gene expression data: a comparative study
447 Citations2008Marcilio CP de Souto, Ivan G. Costa +3 more
This study presents the first large-scale analysis of seven different clustering methods and four proximity measures for the analysis of 35 cancer gene expression data sets and reveals that the finite mixture of Gaussians, followed closely by k-means, exhibited the best performance in terms of recovering the true structure of the data sets.
Journal of the American Statistical AssociationDetecting Features in Spatial Point Processes with Clutter via Model-Based Clustering
415 Citations1998Abhijit Dasgupta, Adrian E. Raftery
This work uses model-based clustering based on a mixture model for the problem of detecting features in spatial point processes when there is substantial clutter, in which features are assumed to generate points according to highly linear multivariate normal densities, and the clutter arises according to a spatial Poisson process.
Computational Statistics & Data AnalysisHigh-dimensional data clustering
367 Citations2007Charles Bouveyron, Stéphane Girard +1 more
A family of Gaussian mixture models designed for high-dimensional data which combine the ideas of dimension reduction and parsimonious modeling give rise to a clustering method based on the Expectation-Maximization algorithm which is called High-Dimensional Data Clustering (HDDC).
Journal of Financial EconometricsThe Generalized Hyperbolic Skew Student's t-Distribution
346 Citations2006Kjersti Aas
Statistics and ComputingParsimonious Gaussian mixture models
331 Citations2008Paul D. McNicholas, Thomas Brendan Murphy
A class of eight parsimonious Gaussian mixture models which are based on the mixtures of factor analyzers model are introduced and the maximum likelihood estimates for the parameters in these models are found using an AECM algorithm.
FEMS Microbiology EcologyDifferences in structure and dynamics of Polynucleobacter communities in a temperate and a subtropical lake, revealed at three phylogenetic levels
271 Citations2006Qinglong L. Wu, Martin W. Hahn
Interhabitat comparisons of the Polynucleobacter community structure in the two lakes in the study, and a previously investigated acidic pond indicated ecological diversification within the phylogenetically narrow Polyn nucleobacter cluster.
Annals of the Institute of Statistical MathematicsThe distribution of the likelihood ratio for mixtures of densities from the one-parameter exponential family
260 Citations1994Dankmar Böhning, Ekkehart Dietz +3 more
BioinformaticsModel-based clustering of microarray expression data via latent Gaussian mixture models
226 Citations2010Paul D. McNicholas, Thomas Brendan Murphy
This modelling approach builds on previous work by introducing a modified factor analysis covariance structure, leading to a family of 12 mixture models, including parsimonious models, which gives very good performance, relative to existing popular clustering techniques, when applied to real gene expression microarray data.
Journal of Multivariate AnalysisMaximum likelihood estimation for multivariate skew normal mixture models
204 Citations2008Tsung‐I Lin
A feasible EM algorithm is developed for finding the maximum likelihood estimates of parameters in this context and a general information-based method for obtaining the asymptotic covariance matrix of themaximum likelihood estimators is presented.
Advances in Data Analysis and ClassificationOn mixtures of skew normal and skew $$t$$ -distributions
190 Citations2013Sharon Lee, Geoffrey J. McLachlan
A systematic classification of the existing skew symmetric distributions into four types is presented, thereby clarifying their close relationships and aiding in understanding the link between some of the proposed expectation-maximization based algorithms for the computation of the maximum likelihood estimates of the parameters of the models.
Canadian Journal of StatisticsA mixture of generalized hyperbolic distributions
188 Citations2015Ryan P. Browne, Paul D. McNicholas
The ability of the authors' models to recover parameters for data from underlying Gaussian and skew‐t distributions is demonstrated and the role of generalized hyperbolic mixtures within the wider model‐based clustering, classification, and density estimation literature is discussed.
IEEE Transactions on Pattern Analysis and Machine IntelligenceMixtures of Shifted AsymmetricLaplace Distributions
168 Citations2014Brian C. Franczak, Ryan P. Browne +1 more
This work marks an important step in the non-Gaussian model-based clustering and classification direction, and a variant of the EM algorithm is developed for parameter estimation by exploiting the relationship with the generalized inverse Gaussian distribution.
IEEE Transactions on Pattern Analysis and Machine IntelligenceMixtures of Factor Analyzers with Common Factor Loadings: Applications to the Clustering and Visualization of High-Dimensional Data
159 Citations2009Jangsun Baek, Geoffrey J. McLachlan +1 more
Pattern Recognition LettersLinear flaw detection in woven textiles using model-based clustering
153 Citations1997Jonathan Campbell, Chris Fraley +2 more
This approach detects a linear pattern in preprocessed images via model-based clustering using an approximate Bayes factor which provides a criterion for assessing the evidence for the presence of a defect.
Computational Statistics & Data AnalysisExtension of the mixture of factor analyzers model to incorporate the multivariate t-distribution
151 Citations2006Geoffrey J. McLachlan, Richard Bean +1 more
An EM-based algorithm is developed for the fitting of mixtures of t-factor analyzers and its application is demonstrated in the clustering of some microarray gene-expression data.
Statistics and ComputingModel-based clustering, classification, and discriminant analysis via mixtures of multivariate t-distributions
143 Citations2011Jeffrey L. Andrews, Paul D. McNicholas
A novel family of mixture models wherein each component is modeled using a multivariate t-distribution with an eigen-decomposed covariance structure is put forth, known as the tEIGEN family.
Journal of Molecular DiagnosticsPrognostic Gene Expression Signatures Can Be Measured in Tissues Collected in RNAlater Preservative
142 Citations2006Dondapati Chowdary, Jessica Lathrop +7 more
This novel finding demonstrates that prognostic signatures can be obtained from RNAlater preservative-suspended tissues, an important step in bringing gene expression signatures to the clinic.
Statistics and ComputingExtending mixtures of multivariate t-factor analyzers
134 Citations2010Jeffrey L. Andrews, Paul D. McNicholas
The extension of the mixtures of multivariate t-factor analyzers model is described to include constraints on the degrees of freedom, the factor loadings, and the error variance matrices to create a family of six mixture models, including parsimonious models.
Statistics and ComputingModel-based clustering with non-elliptically contoured distributions
134 Citations2008Dimitris Karlis, Anais Santourian
Finite mixtures of the normal inverse Gaussian distribution (and its multivariate extensions) are proposed, which start from a density that allows for skewness and fat tails, generalize the existing models, are tractable and have desirable properties.
Computational Statistics & Data AnalysisSerial and parallel implementations of model-based clustering via parsimonious Gaussian mixture models
130 Citations2009Paul D. McNicholas, Thomas Brendan Murphy +2 more
Model-based clustering using a family of Gaussian mixture models, with parsimonious factor analysis-like covariance structure, is described and an ecient algorithm for its implementation is presented, showing its eectiveness when compared to existing software.
Journal of Statistical Planning and InferenceFinite mixtures of multivariate Poisson distributions with application
117 Citations2006Dimitris Karlis, Loukia Meligkotsidou
Computational Statistics & Data AnalysisParsimonious skew mixture models for model-based clustering and classification
96 Citations2013Irene Vrbik, Paul D. McNicholas
Parsimonious skew- t and skew-normal analogues of the GPCM family that employ an eigenvalue decomposition of a scale matrix are introduced and are compared to existing models in both unsupervised and semi-supervised classification frameworks.
Statistics & Probability LettersAnalytic calculations for the EM algorithm for multivariate skew- mixture models
94 Citations2012Irene Vrbik, Paul D. McNicholas
Journal of Statistical Planning and InferenceModel-based classification using latent Gaussian mixture models
89 Citations2009Paul D. McNicholas
A novel model-based classification technique is introduced based on parsimonious Gaussian mixture models that gives excellent classification performance when applied to real food authenticity data on the chemical properties of olive oils from nine areas of Italy.
IEEE Transactions on Pattern Analysis and Machine IntelligenceModel-Based Learning Using a Mixture of Mixtures of Gaussian and Uniform Distributions
78 Citations2011Ryan P. Browne, Paul D. McNicholas +1 more
A mixture model whereby each mixture component is itself a mixture of a multivariate Gaussian distribution and aMultivariate uniform distribution is introduced, which could be used for model-based clustering or model- based classification and focuses on the more general model-Based classification framework.
Journal of Multivariate AnalysisExtending mixtures of factor models using the restricted multivariate skew-normal distribution
78 Citations2015Tsung‐I Lin, Geoffrey J. McLachlan +1 more
The proposed MSNFA model provides an approach to model-based density estimation and clustering of high-dimensional data exhibiting asymmetric characteristics and a computationally feasible Expectation Conditional Maximization (ECM) algorithm is developed for computing the maximum likelihood estimates of model parameters.
BioinformaticsMixtures of common <i>t</i>-factor analyzers for clustering high-dimensional microarray data
74 Citations2011Jangsun Baek, Geoffrey J. McLachlan
An EM algorithm is developed for the fitting of mixtures of common t-factor analyzers that can handle data with tails longer than that of the normal distribution, is robust against outliers and allows the data to be displayed in low-dimensional plots.
Statistics and ComputingDimension reduction for model-based clustering
65 Citations2009Luca Scrucca
A dimension reduction method for visualizing the clustering structure obtained from a finite mixture of Gaussian densities, which aims at reducing the dimensionality by identifying a set of linear combinations of the original features which capture most of the cluster structure contained in the data.
Journal of Statistical Planning and InferenceClustering gene expression time course data using mixtures of multivariate t-distributions
57 Citations2011Paul D. McNicholas, Sanjeena Subedi
A very general and flexible model-based technique is used to cluster longitudinal data, usingixtures of multivariate t-distributions with a linear model for the mean and a modified Cholesky-decomposed covariance structure.
BioinformaticsArrayCluster: an analytic tool for clustering, data visualization and module finder on gene expression profiles
40 Citations2006Ryo Yoshida, Tomoyuki Higuchi +2 more
The ArrayCluster is a freely available software to perform the mixed factors analysis and provides some analytic tools for clustering DNA microarray experiments, data visualization and an automatic detector for module transcriptional of genes that are relevant to the calibrated molecular subtypes and so on.
Proceedings. 2004 IEEE Computational Systems Bioinformatics Conference, 2004. CSB 2004.A mixed factors model for dimension reduction and extraction of a group structure in gene expression data
39 Citations2004Ryo Yoshida, Tomoyuki Higuchi +1 more
This work attempts a methodological extension of the factor analysis to prevent from the occurrence of overfilling, but also to handle the issues of clustering, data compression and extracting a set of genes to be relevant to explain the group structure.
Computational StatisticsSupervised learning of multivariate skew normal mixture models with missing information
35 Citations2009Tzy-Chy Lin, Tsung‐I Lin +2 more
The proposed mixture analyzer, including the most commonly used Gaussian mixtures as a special case, allows practitioners to handle incomplete multivariate data sets in a wide range of considerations.
Statistics & Probability LettersDimension reduction for model-based clustering via mixtures of shifted asymmetric Laplace distributions
26 Citations2013Katherine Morris, Paul D. McNicholas
A dimension reduction method for model-based clustering via a finite mixture of shifted asymmetric Laplace distributions is introduced, based on existing work within the Gaussian paradigm and relies on identification of a reduced subspace.
Advances in Data Analysis and ClassificationDimension reduction for model-based clustering via mixtures of multivariate $$t$$ t -distributions
19 Citations2013Katherine Morris, Paul D. McNicholas +1 more
A dimension reduction method for model-based clustering obtained from a finite mixture of $$t$$t-distributions" based on existing work on reducing dimensionality in the case of finite Gaussian mixtures is introduced.
Journal of Multivariate AnalysisMixtures of common factor analyzers for high-dimensional data with missing information
18 Citations2013Wan‐Lun Wang
A computationally flexible EM-type algorithm for parameter estimation of the MCFA model with partially observed data is established, and two auxiliary permutation matrices are incorporated into the estimating procedure for exactly extracting the location of observed and missing components of each observation.
arXiv (Cornell University)Mixtures of Variance-Gamma Distributions
17 Citations2013Sharon M. McNicholas, Paul D. McNicholas +1 more
The proposed mixture of variance-gamma distributions is a special case of the recently developed mixture of generalized hyperbolic distributions, and a restriction is required to ensure identifiability, and serves as an alternative to the ubiquitous mixture of Gaussian distributions.
arXiv (Cornell University)Parsimonious Shifted Asymmetric Laplace Mixtures
16 Citations2013Brian C. Franczak, Paul D. McNicholas +2 more
A family of parsimonious shifted asymmetric Laplace mixture models is introduced, and this novel family of models is applied to real data, where it is compared to its Gaussian analogue within clustering and classication paradigms.