/D-CCA

A Decomposition-based Canonical Correlation Analysis for High-dimensional Datasets (JASA-20 paper)

Primary LanguagePython

D-CCA: A Decomposition-based Canonical Correlation Analysis for High-dimensional Datasets

This python package implements the D-CCA method proposed in [1] for K=2 datasets. See example.py for details, with Python 3.5 or above. For K>2 datasets, please use the D-GCCA method.

D-CCA conducts the following decomposition:

for

where and share the same latent factors, but and have uncorrelated latent factors.

Note that should be row-mean centered.

Please cite the article [1] for this package, which is available here.

[1] Hai Shu, Xiao Wang & Hongtu Zhu (2020) D-CCA: A Decomposition-based Canonical Correlation Analysis for High-dimensional Datasets. Journal of the American Statistical Association, 115(529): 292-306. DOI: 10.1080/01621459.2018.1543599