Extension of the mixture of factor analyzers model to incorporate the multivariate t-distribution
Computational Statistics & Data AnalysisPublished 11 October 2006
Geoffrey J. McLachlan, Richard Bean, L. Ben-Tovim Jones
Citations151
SJR quartileQ1
SJR score0.89
SNIP1.38
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
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.
Abstract
Mixtures of factor analyzers enable model-based density estimation to be undertaken for high-dimensional data, where the number of observations n is small relative to their dimension p. However, this approach is sensitive to outliers as it is based on a mixture model in which the multivariate normal family of distributions is assumed for the component error and factor distributions. An extension to mixtures of t-factor analyzers is considered, whereby the multivariate t-family is adopted for the component error and factor distributions. An EM-based algorithm is developed for the fitting of mixtures of t-factor analyzers. Its application is demonstrated in the clustering of some microarray gene-expression data.
Keywords
Computer ScienceMathematicsBiochemistry, Genetics and Molecular Biology
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
NatureMolecular portraits of human breast tumours
16,304 Citations2000Charles M. Perou, Thérese Sørlie +16 more
Variation in gene expression patterns in a set of 65 surgical specimens of human breast tumours from 42 different individuals were characterized using complementary DNA microarrays representing 8,102 human genes, providing a distinctive molecular portrait of each tumour.
NatureGene expression profiling predicts clinical outcome of breast cancer
9,608 Citations2002Laura van ‘t Veer, Hongyue Dai +14 more
DNA microarray analysis on primary breast tumours of 117 young patients is used and supervised classification is applied to identify a gene expression signature strongly predictive of a short interval to distant metastases (‘poor prognosis’ signature) in patients without tumour cells in local lymph nodes at diagnosis, providing a strategy to select patients who would benefit from adjuvant therapy.
Wiley series in probability and statisticsFinite Mixture Models
7,410 Citations2018Geoffrey J. McLachlan, David Peel +1 more
The aim of this article is to provide an up-to-date account of the theory and methodological developments underlying the applications of finite mixture models.
New England Journal of MedicineA Gene-Expression Signature as a Predictor of Survival in Breast Cancer
6,520 Citations2002Marc J. van de Vijver, Yudong D. He +19 more
The gene-expression profile studied is a more powerful predictor of the outcome of disease in young patients with breast cancer than standard systems based on clinical and histologic criteria.
TechnometricsThe EM Algorithm and Extensions
5,108 Citations1998Debashis Kushary, Geoffrey J. McLachlan +1 more
Journal of the American Statistical AssociationStatistical Analysis of Finite Mixture Distributions.
2,940 Citations1987Bruce G. Lindsay, D. M. Titterington +2 more
BiometricsModel-Based Gaussian and Non-Gaussian Clustering
2,350 Citations1993Jeffrey D. Banfield, Adrian E. Raftery
The classification maximum likelihood approach is sufficiently general to encompass many current clustering algorithms, including those based on the sum of squares criterion and on the criterion of Friedman and Rubin (1967), but it is restricted to Gaussian distributions and it does not allow for noise.
Queensland's institutional digital repository (The University of Queensland)Mixture models : inference and applications to clustering
2,074 Citations1988Geoffrey J. McLachlan, K. E. Basford
Neural ComputationMixtures of Probabilistic Principal Component Analyzers
1,903 Citations1999Michael E. Tipping, Chris Bishop
PCA is formulated within a maximum likelihood framework, based on a specific form of gaussian latent variable model, which leads to a well-defined mixture model for probabilistic principal component analyzers, whose parameters can be determined using an expectation-maximization algorithm.
BiometrikaMaximum likelihood estimation via the ECM algorithm: A general framework
1,814 Citations1993Xiao‐Li Meng, Donald B. Rubin
Encyclopedia of Statistics in Behavioral ScienceFinite Mixture Distributions
1,287 Citations2005Barry J. Everitt
Statistics and ComputingRobust mixture modelling using the t distribution
930 Citations2000David Peel, Geoffrey J. McLachlan
The use of the ECM algorithm to fit this t mixture model is described and examples of its use are given in the context of clustering multivariate data in the presence of atypical observations in the form of background noise.
Cambridge University Press eBooksMultivariate T-Distributions and Their Applications
917 Citations2004Samuel Kotz, Saralees Nadarajah
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.
BioinformaticsA mixture model-based approach to the clustering of microarray expression data
564 Citations2002Geoffrey J. McLachlan, Richard Bean +1 more
The usefulness of the EMMIX-GENE approach for the clustering of tissue samples is demonstrated on two well-known data sets on colon and leukaemia tissues, and relevant subsets of the genes are able to be selected that reveal interesting clusterings of the tissues that are either consistent with the external classified tissues or with background and biological knowledge of these sets.
Proceedings of the National Academy of SciencesGene expression profiles of primary breast tumors maintained in distant metastases
431 Citations2003Britta Weigelt, Annuska M. Glas +4 more
By gene expression profiling, it is shown that human primary breast tumors are strikingly similar to the distant metastases of the same patient, suggesting that metastatic capability in breast cancer is an inherent feature and is not based on clonal selection.
The Annals of StatisticsA Constrained Formulation of Maximum-Likelihood Estimation for Normal Mixture Distributions
398 Citations1985Richard J. Hathaway
Computational Statistics & Data AnalysisModelling high-dimensional data by mixtures of factor analyzers
271 Citations2003Geoffrey J. McLachlan, David Peel +1 more
This work focuses on mixtures of factor analyzers from the perspective of a method for model-based density estimation from high-dimensional data, and hence for the clustering of such data.
NeurocomputingRobust Bayesian mixture modelling
248 Citations2005Markus Svensén, Chris Bishop
A Bayesian approach to mixture modelling based on Student-t distributions, which are heavier tailed than Gaussians and hence more robust, is developed, which includes Gaussian mixtures as a special case.
Medical Entomology and ZoologyComputer-Assisted Analysis of Mixtures and Applications: Meta-Analysis, Disease Mapping, and Others
228 Citations1999Dankmar Böhning
Theory of nonparametric mixture models algorithms the likelihood ratio test for the number of components C.A.MAN-applications and meta-analysis moment estimators of the variance of the mixing distribution are studied.
Wiley series in probability and statisticsMixtures of Factor Analyzers
200 Citations2000Geoffrey J. McLachlan, David Peel
Latent Variable Models
178 Citations1998Chris Bishop
This chapter provides an overview of latent variable models for representing continuous variables and shows how a particular form of linear latent variable model can be used to provide a probabilistic formulation of the well-known technique of principal components analysis (PCA).
Lecture notes in computer scienceRobust cluster analysis via mixtures of multivariate t-distributions
162 Citations1998Geoffrey J. McLachlan, David Peel
The expectation-maximization (EM) algorithm can be used to fit mixtures of multivariate t-distributions by maximum likelihood and it is demonstrated how the use of t-components provides less extreme estimates of the posterior probabilities of cluster membership.
The Annals of StatisticsBreakdown points for maximum likelihood estimators of location–scale mixtures
106 Citations2004Christian Hennig
It turns out that the two alternatives, while adding stability in the presence of outliers of moderate size, do not possess a substantially better breakdown behavior than estimation based on Normal mixtures.
Statistical Methods & ApplicationsA likelihood-based constrained algorithm for multivariate normal mixture models
85 Citations2004Salvatore Ingrassia
Machine LearningMixtures of Factor Analysers. Bayesian Estimation and Inference by Stochastic Simulation
74 Citations2003Ernest Fokoué, D. M. Titterington
A Bayesian approach toFactor Analysis is adopted, and more specifically a treatment that bases estimation and inference on the stochastic simulation of the posterior distributions of interest, and can be envisaged as an alternative to the other approaches used for this model.
NeurocomputingProbabilistic PCA for t distributions
41 Citations2006Jianhua Zhao, Quan Jiang
A new dimension reduction tool is theoretically a generalization of the above probabilistic PCA; in practice, it requires smaller model size and has better performance as shown by the experimental results.
Computational Statistics & Data AnalysisOn computing the largest fraction of missing information for the EM algorithm and the worst linear function for data augmentation
37 Citations1999Daniel Coleman
Results of simulation studies on multivariate normal data are given showing that the power method for eigencomputation becomes more efficient as the data dimension increases than methods that use a finite-difference approximation to the Jacobian, which is the only general-purpose alternative available.
Mixtures of Probabilistic Principal Component Analysers
22 Citations1998Michael E. Tipping, Cambridge Cb Nh +1 more
Computational Statistics & Data AnalysisSome computational issues in cluster analysis with no a priori metric
21 Citations1999Daniel Coleman, Xioapeng Dong +3 more