Dimensionality Reduction of Clustered Data Sets

Author

Sanguinetti, Guido

Author_Institution

Univ. of Sheffield, Sheffield

Volume

30

Issue

3

fYear

2008

fDate

3/1/2008 12:00:00 AM

Firstpage

535

Lastpage

540

Abstract

We present a novel probabilistic latent variable model to perform linear dimensionality reduction on data sets which contain clusters. We prove that the maximum likelihood solution of the model is an unsupervised generalization of linear discriminant analysis. This provides a completely new approach to one of the most established and widely used classification algorithms. The performance of the model is then demonstrated on a number of real and artificial data sets.

Keywords

maximum likelihood estimation; pattern classification; pattern clustering; classification algorithms; clustered data sets; linear dimensionality reduction; linear discriminant analysis; maximum likelihood solution; probabilistic latent variable model; Algorithm design and analysis; Bioinformatics; Classification algorithms; Clustering algorithms; Computer vision; Feature extraction; Independent component analysis; Linear discriminant analysis; Machine learning algorithms; Principal component analysis; clustering; dimensionality reduction; discriminant analysis; probabilistic algorithms; Algorithms; Artificial Intelligence; Cluster Analysis; Databases, Factual; Information Storage and Retrieval; Numerical Analysis, Computer-Assisted; Pattern Recognition, Automated; Reproducibility of Results; Sensitivity and Specificity;

fLanguage

English

Journal_Title

Pattern Analysis and Machine Intelligence, IEEE Transactions on

Publisher

ieee

ISSN

0162-8828

Type

jour

DOI

10.1109/TPAMI.2007.70819

Filename

4378396