Bayesian Group Factor Analysis

Seppo Virtanen, Arto Klami, Suleiman Khan, Samuel Kaski
Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, PMLR 22:1269-1277, 2012.

Abstract

We introduce a factor analysis model that summarizes the dependencies between observed variable groups, instead of dependencies between individual variables as standard factor analysis does. A group may correspond to one view of the same set of objects, one of many data sets tied by co-occurrence, or a set of alternative variables collected from statistics tables to measure one property of interest. We show that by assuming group-wise sparse factors, active in a subset of the sets, the variation can be decomposed into factors explaining relationships between the sets and factors explaining away set-specific variation. We formulate the assumptions in a Bayesian model providing the factors, and apply the model to two data analysis tasks, in neuroimaging and chemical systems biology.

Cite this Paper

BibTeX
@InProceedings{pmlr-v22-virtanen12, title = {Bayesian Group Factor Analysis}, author = {Virtanen, Seppo and Klami, Arto and Khan, Suleiman and Kaski, Samuel}, booktitle = {Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics}, pages = {1269--1277}, year = {2012}, editor = {Lawrence, Neil D. and Girolami, Mark}, volume = {22}, series = {Proceedings of Machine Learning Research}, address = {La Palma, Canary Islands}, month = {21--23 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v22/virtanen12/virtanen12.pdf}, url = {https://proceedings.mlr.press/v22/virtanen12.html}, abstract = {We introduce a factor analysis model that summarizes the dependencies between observed variable groups, instead of dependencies between individual variables as standard factor analysis does. A group may correspond to one view of the same set of objects, one of many data sets tied by co-occurrence, or a set of alternative variables collected from statistics tables to measure one property of interest. We show that by assuming group-wise sparse factors, active in a subset of the sets, the variation can be decomposed into factors explaining relationships between the sets and factors explaining away set-specific variation. We formulate the assumptions in a Bayesian model providing the factors, and apply the model to two data analysis tasks, in neuroimaging and chemical systems biology.} }
Endnote
%0 Conference Paper %T Bayesian Group Factor Analysis %A Seppo Virtanen %A Arto Klami %A Suleiman Khan %A Samuel Kaski %B Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2012 %E Neil D. Lawrence %E Mark Girolami %F pmlr-v22-virtanen12 %I PMLR %P 1269--1277 %U https://proceedings.mlr.press/v22/virtanen12.html %V 22 %X We introduce a factor analysis model that summarizes the dependencies between observed variable groups, instead of dependencies between individual variables as standard factor analysis does. A group may correspond to one view of the same set of objects, one of many data sets tied by co-occurrence, or a set of alternative variables collected from statistics tables to measure one property of interest. We show that by assuming group-wise sparse factors, active in a subset of the sets, the variation can be decomposed into factors explaining relationships between the sets and factors explaining away set-specific variation. We formulate the assumptions in a Bayesian model providing the factors, and apply the model to two data analysis tasks, in neuroimaging and chemical systems biology.
RIS
TY - CPAPER TI - Bayesian Group Factor Analysis AU - Seppo Virtanen AU - Arto Klami AU - Suleiman Khan AU - Samuel Kaski BT - Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics DA - 2012/03/21 ED - Neil D. Lawrence ED - Mark Girolami ID - pmlr-v22-virtanen12 PB - PMLR DP - Proceedings of Machine Learning Research VL - 22 SP - 1269 EP - 1277 L1 - http://proceedings.mlr.press/v22/virtanen12/virtanen12.pdf UR - https://proceedings.mlr.press/v22/virtanen12.html AB - We introduce a factor analysis model that summarizes the dependencies between observed variable groups, instead of dependencies between individual variables as standard factor analysis does. A group may correspond to one view of the same set of objects, one of many data sets tied by co-occurrence, or a set of alternative variables collected from statistics tables to measure one property of interest. We show that by assuming group-wise sparse factors, active in a subset of the sets, the variation can be decomposed into factors explaining relationships between the sets and factors explaining away set-specific variation. We formulate the assumptions in a Bayesian model providing the factors, and apply the model to two data analysis tasks, in neuroimaging and chemical systems biology. ER -
APA
Virtanen, S., Klami, A., Khan, S. & Kaski, S.. (2012). Bayesian Group Factor Analysis. Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 22:1269-1277 Available from https://proceedings.mlr.press/v22/virtanen12.html.

Related Material