Heterogeneous component analysis

Shigeyuki Oba, Motoaki Kawanabe, Klaus Muller, Shin Ishii

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In bioinformatics it is often desirable to combine data from various measurement sources and thus structured feature vectors are to be analyzed that possess different intrinsic blocking characteristics (e.g., different patterns of missing values, observation noise levels, effective intrinsic dimensionalities). We propose a new machine learning tool, heterogeneous component analysis (HCA), for feature extraction in order to better understand the factors that underlie such complex structured heterogeneous data. HCA is a linear block-wise sparse Bayesian PCA based not only on a probabilistic model with block-wise residual variance terms but also on a Bayesian treatment of a block-wise sparse factor-loading matrix. We study various algorithms that implement our HCA concept extracting sparse heterogeneous structure by obtaining common components for the blocks and specific components within each block. Simulations on toy and bioinformatics data underline the usefulness of the proposed structured matrix factorization concept.

Original languageEnglish
Title of host publicationAdvances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference
Publication statusPublished - 2009 Dec 1
Externally publishedYes
Event21st Annual Conference on Neural Information Processing Systems, NIPS 2007 - Vancouver, BC, Canada
Duration: 2007 Dec 32007 Dec 6

Other

Other21st Annual Conference on Neural Information Processing Systems, NIPS 2007
CountryCanada
CityVancouver, BC
Period07/12/307/12/6

Fingerprint

Bioinformatics
Factorization
Learning systems
Feature extraction
Statistical Models

ASJC Scopus subject areas

  • Information Systems

Cite this

Oba, S., Kawanabe, M., Muller, K., & Ishii, S. (2009). Heterogeneous component analysis. In Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference

Heterogeneous component analysis. / Oba, Shigeyuki; Kawanabe, Motoaki; Muller, Klaus; Ishii, Shin.

Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference. 2009.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Oba, S, Kawanabe, M, Muller, K & Ishii, S 2009, Heterogeneous component analysis. in Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference. 21st Annual Conference on Neural Information Processing Systems, NIPS 2007, Vancouver, BC, Canada, 07/12/3.
Oba S, Kawanabe M, Muller K, Ishii S. Heterogeneous component analysis. In Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference. 2009
Oba, Shigeyuki ; Kawanabe, Motoaki ; Muller, Klaus ; Ishii, Shin. / Heterogeneous component analysis. Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference. 2009.
@inproceedings{ad0360ae56324399a7d331ee7bc4f2d4,
title = "Heterogeneous component analysis",
abstract = "In bioinformatics it is often desirable to combine data from various measurement sources and thus structured feature vectors are to be analyzed that possess different intrinsic blocking characteristics (e.g., different patterns of missing values, observation noise levels, effective intrinsic dimensionalities). We propose a new machine learning tool, heterogeneous component analysis (HCA), for feature extraction in order to better understand the factors that underlie such complex structured heterogeneous data. HCA is a linear block-wise sparse Bayesian PCA based not only on a probabilistic model with block-wise residual variance terms but also on a Bayesian treatment of a block-wise sparse factor-loading matrix. We study various algorithms that implement our HCA concept extracting sparse heterogeneous structure by obtaining common components for the blocks and specific components within each block. Simulations on toy and bioinformatics data underline the usefulness of the proposed structured matrix factorization concept.",
author = "Shigeyuki Oba and Motoaki Kawanabe and Klaus Muller and Shin Ishii",
year = "2009",
month = "12",
day = "1",
language = "English",
isbn = "160560352X",
booktitle = "Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference",

}

TY - GEN

T1 - Heterogeneous component analysis

AU - Oba, Shigeyuki

AU - Kawanabe, Motoaki

AU - Muller, Klaus

AU - Ishii, Shin

PY - 2009/12/1

Y1 - 2009/12/1

N2 - In bioinformatics it is often desirable to combine data from various measurement sources and thus structured feature vectors are to be analyzed that possess different intrinsic blocking characteristics (e.g., different patterns of missing values, observation noise levels, effective intrinsic dimensionalities). We propose a new machine learning tool, heterogeneous component analysis (HCA), for feature extraction in order to better understand the factors that underlie such complex structured heterogeneous data. HCA is a linear block-wise sparse Bayesian PCA based not only on a probabilistic model with block-wise residual variance terms but also on a Bayesian treatment of a block-wise sparse factor-loading matrix. We study various algorithms that implement our HCA concept extracting sparse heterogeneous structure by obtaining common components for the blocks and specific components within each block. Simulations on toy and bioinformatics data underline the usefulness of the proposed structured matrix factorization concept.

AB - In bioinformatics it is often desirable to combine data from various measurement sources and thus structured feature vectors are to be analyzed that possess different intrinsic blocking characteristics (e.g., different patterns of missing values, observation noise levels, effective intrinsic dimensionalities). We propose a new machine learning tool, heterogeneous component analysis (HCA), for feature extraction in order to better understand the factors that underlie such complex structured heterogeneous data. HCA is a linear block-wise sparse Bayesian PCA based not only on a probabilistic model with block-wise residual variance terms but also on a Bayesian treatment of a block-wise sparse factor-loading matrix. We study various algorithms that implement our HCA concept extracting sparse heterogeneous structure by obtaining common components for the blocks and specific components within each block. Simulations on toy and bioinformatics data underline the usefulness of the proposed structured matrix factorization concept.

UR - http://www.scopus.com/inward/record.url?scp=84889725544&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889725544&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84889725544

SN - 160560352X

SN - 9781605603520

BT - Advances in Neural Information Processing Systems 20 - Proceedings of the 2007 Conference

ER -