TY - GEN
T1 - Nonparametric factor analysis with beta process priors
AU - Paisley, John
AU - Carin, Lawrence
N1 - Generated from Scopus record by KAUST IRTS on 2021-02-09
PY - 2009/12/9
Y1 - 2009/12/9
N2 - We propose a nonparametric extension to the factor analysis problem using a beta process prior. This beta process factor analysis (BP-FA) model allows for a dataset to be decomposed into a linear combination of a sparse set of factors, providing information on the underlying structure of the observations. As with the Dirichlet process, the beta process is a fully Bayesian conjugate prior, which allows for analytical posterior calculation and straightforward inference. We derive a variational Bayes inference algorithm and demonstrate the model on the MNIST digits and HGDP-CEPH cell line panel datasets.
AB - We propose a nonparametric extension to the factor analysis problem using a beta process prior. This beta process factor analysis (BP-FA) model allows for a dataset to be decomposed into a linear combination of a sparse set of factors, providing information on the underlying structure of the observations. As with the Dirichlet process, the beta process is a fully Bayesian conjugate prior, which allows for analytical posterior calculation and straightforward inference. We derive a variational Bayes inference algorithm and demonstrate the model on the MNIST digits and HGDP-CEPH cell line panel datasets.
UR - http://portal.acm.org/citation.cfm?doid=1553374.1553474
UR - http://www.scopus.com/inward/record.url?scp=70049097438&partnerID=8YFLogxK
U2 - 10.1145/1553374.1553474
DO - 10.1145/1553374.1553474
M3 - Conference contribution
SN - 9781605585161
SP - 777
EP - 784
BT - Proceedings of the 26th International Conference On Machine Learning, ICML 2009
ER -