A generalization of principal component analysis to the exponential family

by: M Collins, S Dasgupta, R Schapire

(2001)

View FullText article

CiteSeer (at Penn State), CiteSeer (at MIT), CiteSeer (at Zurich), CiteSeer (at Singapore)

Reviews [Write a review of this article]

There are no reviews of this article

Find related articles from these CiteULike users

vlachmore, mdreid, Statistical Machine Learning (group), tdhollan

Find related articles with these CiteULike tags

analysis, bregman_divergence, component, dimensionality, dimensional_reduction, exponential, exponential_family, family, pca, principal, reduction

Abstract

Principal component analysis (PCA) is a commonly applied technique for dimensionality reduction. PCA implicitly minimizes a squared loss function, which may be inappropriate for data that is not real-valued, such as binary-valued data. This paper draws on ideas from the Exponential family, Generalized linear models, and Bregman distances, to give a generalization of PCA to loss functions that we argue are better suited to other data types. We describe algorithms for minimizing the loss...

@misc{citeulike:2891300, abstract = {Principal component analysis (PCA) is a commonly applied technique for dimensionality reduction. PCA implicitly minimizes a squared loss function, which may be inappropriate for data that is not real-valued, such as binary-valued data. This paper draws on ideas from the Exponential family, Generalized linear models, and Bregman distances, to give a generalization of PCA to loss functions that we argue are better suited to other data types. We describe algorithms for minimizing the loss...}, author = {Collins, M. and Dasgupta, S. and Schapire, R. }, citeulike-article-id = {2891300}, keywords = {bregman\_divergence, dimensional\_reduction, exponential\_family, pca}, posted-at = {2008-06-17 13:57:03}, priority = {2}, title = {A generalization of principal component analysis to the exponential family}, url = {http://citeseer.ist.psu.edu/542751.html}, year = {2001} }

RIS record

TY - GEN ID - citeulike:2891300 L3 - citeulike-article-id:2891300 TI - A generalization of principal component analysis to the exponential family N2 - Principal component analysis (PCA) is a commonly applied technique for dimensionality reduction. PCA implicitly minimizes a squared loss function, which may be inappropriate for data that is not real-valued, such as binary-valued data. This paper draws on ideas from the Exponential family, Generalized linear models, and Bregman distances, to give a generalization of PCA to loss functions that we argue are better suited to other data types. We describe algorithms for minimizing the loss... KW - bregman_divergence KW - dimensional_reduction KW - exponential_family KW - pca AU - Collins, M AU - Dasgupta, S AU - Schapire, R PY - 2001/// UR - http://citeseer.ist.psu.edu/542751.html ER -

RIS BibTeX