Links between PPCA and subspace methods for complete Gaussian density estimation

Author

Chong Wang ; Wenyuan Wang

Author_Institution

Dept. of Autom., Tsinghua Univ., Beijing

Volume

17

Issue

3

fYear

2006

fDate

5/1/2006 12:00:00 AM

Firstpage

789

Lastpage

792

Abstract

High-dimensional density estimation is a fundamental problem in pattern recognition and machine learning areas. In this letter, we show that, for complete high-dimensional Gaussian density estimation, two widely used methods, probabilistic principal component analysis and a typical subspace method using eigenspace decomposition, actually give the same results. Additionally, we present a unified view from the aspect of robust estimation of the covariance matrix

Keywords

Gaussian processes; covariance matrices; eigenvalues and eigenfunctions; estimation theory; learning (artificial intelligence); pattern recognition; principal component analysis; complete high-dimensional Gaussian density estimation; covariance matrix; eigenspace decomposition; machine learning; pattern recognition; probabilistic principal component analysis; robust estimation; subspace methods; Automation; Covariance matrix; Eigenvalues and eigenfunctions; Gaussian noise; Information processing; Machine learning; Matrix decomposition; Noise robustness; Pattern recognition; Principal component analysis; Complete Gaussian density estimation; eigenspace decomposition; probabilistic principal component analysis (PPCA); subspace method; Algorithms; Artificial Intelligence; Computer Simulation; Information Storage and Retrieval; Models, Statistical; Neural Networks (Computer); Normal Distribution; Pattern Recognition, Automated; Principal Component Analysis; Systems Theory;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/TNN.2006.871718

Filename

1629099