The Bayesian CCA model as implemented here was originally presented by Virtanen et al. (2011), but a more comprehensive treatment is found in Klami et al. (2013). The latter also explains the BIBFA model. The GFA extends CCA to multiple data sources (or groups of variables), providing interpretable linear factorizations that describe variation shared by all possible subsets of sources. It was originally presented by Virtanen et al. (2012). Later Klami et al. (2014) provide a more extensive literature review and present a novel hierarchical low-rank ARD prior for the factor loadings to better account for inter-source relationships.
We recommend that scientific publications using the code for CCA or BIBFA cite Klami et al. (2013), and publications using the code for GFA cite Virtanen et al. (2012), until Klami et al. (2014) has been published.
The package is based on the research done in the SMLB group, Helsinki Institute for Information Technology HIIT, Department of Information and Computer Science, Aalto University, http://research.ics.aalto.fi/mi/.
| Package: |
| CCAGFA |
| Type: |
| Package |
| Version: |
| 1.0.4 |
| Date: |
| 2013-04-23 |
| License: |
| GPL (>= 2) |
Virtanen, S. and Klami, A., and Khan, S.A. and Kaski, S.: Baysian group factor analysis. In Proceedings of the 15th International Conference on Artificial Intelligence and Statistics (AISTATS), volume 22 of JMLR W&CP, pages 1269-1277, 2012.
Klami, A. and Virtanen, S., and Kaski, S.: Bayesian Canonical Correlation Analysis. Journal of Machine Learning Research,14:965-1003, 2013.
Klami, A. and Virtanen, S., Leppaaho, E., and Kaski, S.: Group Factor Analysis. IEEE Transactions on Neural Networks and Learning Systems, to appear.
# require(CCAGFA)
# demo(CCAGFAexample)
Run the code above in your browser using DataLab